Introduction
MRD monitoring is recommended for treatment response evaluation in clinical AML trials (ELN-2022 guidelines: Heuser et al, PMID 34724563). Besides molecular biology methods, multiparametric flow cytometry represents the most reliable approach. In this study, we evaluated the impact of flow MRD in patients (pts) enrolled in the Backbone Intergroup BIG-1 trial (NCT02416388) from the Acute Leukemia French Intergroup (ALFA-FILO) between 2018 and 2021. For that purpose, a standardized multicentric MRD flow approach based on ELN recommendations and combining LAIP/Dfn and CD34+CD38- Leukemic Stem Cell (LSC) detection as previously reported (Zeijlemaker et al, PMID 30542144; Ngai et al, PMID 36898087) was implemented across 30 French hematology laboratories.
Methods
Multiparametric flow cytometry was performed on fresh bone marrow (BM) sample using a 2-tubes panel, with minimal mandatory 8-color markers by tube, to characterize the pattern of LAIP (Leukemia Aberrant ImmunoPhenotype)/Dfn (Different from normal) and CD34+CD38- LSCs. Our network has established recommendations from Wet labs procedure to clinical reports using common panels, flow cytometers settings and analysis strategy (Plesa et al, Annual ASH Meeting 2022). A « backbone » using CD34/CD38/CD45/CD117 was used in both tubes, completed by CD7, CD56, CD13, CD33, HLADR, CD19 for the 1st tube and CD90Thy-1, Mix (CD97/CLL1/TIM3), CD45RA, CD123 for the 2nd one. A total of at least 500,000 to one million cells was acquired in each tube. An Assurance Quality Program was implemented in all laboratories during the study. All data were reviewed independently by the MRDflow network coordinator.
Results
Between 2018 and 2021, 315/1228 pts treated in the BIG-1 trial (median age, 49 years; ELN-2022 risk: 97 favorable, 87 intermediate, 110 adverse, 21 unknown) were studied by flow. No differences were observed between pts with or w/o flow evaluation according to main patient/AML characteristics and outcomes.
At diagnosis, using a 1% of CD45+/ssc BM blast cells cut-off, 88/280 pts (31.4%) were LSC-high and 192/280 pts (68.6%) were LSC-low, with a strong correlation between LSC-high incidence and ELN-2022 risk (6.2, 37.3 and 45.6% in the favorable, intermediate and adverse group, respectively; p<0.001). As expected, OS was thus significantly lower in the LSC-high group (3y-OS, 45% [95% confidence interval, 35-56] vs 69% [62-75]; p<0.001).
Using a 0.1% of CD45+ BM cells cut-off, the numbers of pts with positive LAIP/Dfn MRD was 90/298 (30.2%) at post-induction (MRD1) and 69/253 (27.3%) at post-consolidation (MRD2) timepoint. For both timepoints, OS was significantly lower in LAIP/Dfn-positive vs LAIP/Dfn-negative pts (3y-OS, 50% [39-60] vs 75% [68-80] for MRD1 and 49% [37-60] vs 78% [71-83] for MRD2; p<0.001 for both).
Using a 0.01% of CD45+ BM cells cut-off, the numbers of pts with positive LSC MRD were 57/222 (25.7%) at MRD1 and 38/197 (19.3%) at MRD2. Again, OS was significantly lower in LSC-positive vs LSC-negative pts at both timepoints (3y-OS, 47% [33-59] vs 77% [69-82] for MRD1 and 42% [26-57] vs 80% [73-86] for MRD2; p<0.001 for both).
Finally, the “scoring system” was validated when combining both LAIP/Dfn and LSC markers at MRD1 and MRD2. Among 220 pts tested for both markers at MRD1, the LAIP-Dfn/LSC score was -/-, +/-, -/+ and +/+ in 135, 24, 33 and 28 pts, respectively. At 3 years, OS estimates were 78% (70-84), 68% (47-82), 71% (48-85) and 30% (16-46), in these 4 subgroups, respectively (p=0.033).
Among 193 pts tested for both markers at MRD2, the LAIP-Dfn/LSC score was -/-, +/-, -/+ and +/+ in 125, 12, 26 and 30 pts, respectively. At 3 years, OS estimates were 82% (74-88), 70% (50-83), 75% (41-91) and 27% (12-44), in these 4 subgroups, respectively (p=0.007).
Similar predictive values were observed when censoring patients allografted in first remission at transplant date.
Finally, the predictive values of LAIP/Dfn at MRD1 and MRD2, LSC at MRD1 and MRD2, and combined score at MRD2 (but not at MRD1) remained significant after adjustment on the ELN-2022 risk.
Conclusions
Our first experience demonstrates the feasibility of a multicentric flow MRD approach in AML. We confirm in the prospective French BIG-1 trial that flow MRD monitoring combining LAIP/Dfn and CD34+CD38- LSC allows better defining AML patient prognosis independently of the ELN-2022 risk.
Recent Comments