Clinical Relevance of– Limit of Detection (LOD) – Limit of Quantification (LOQ) - Based Flow Cytometry Approach for Measurable Residual Disease (MRD) Assessment in Acute Myeloid Leukemia (AML)

Background: In Acute Myeloid Leukemia (AML), identification of measurable residual disease (MRD) thresholds with clinical significance is still a matter of debate. For this purpose, multiparametric flow cytometry (MFC) is extensively employed for MRD quantification, due to high sensitivity (down to 1:10^-3/10^-5 cells) and wide applicability (up to 90% of cases). The identification of 20 clustered residual leukemic cells seems sufficient for the recognition of MRD presence (lower limit of detection [LOD]), whereas a cluster of 50 events may be the minimum threshold for the quantification of a cell population (lower limit of quantitation [LOQ]), provided a sufficient denominator of relevant events (500’000-1’000’000) is acquired.

Methods: Using a MFC assay, we assessed the predictive power of a threshold calculated applying the criteria of LOD and LOQ on 261 intensively treated AML patients enrolled in the GIMEMA AML1310 prospective trial.According to the protocol design, patients with a bone marrow residual leukemic cells count (RLCc) equal or above 0.035% of the total no. of mononuclear (MNC) cells qualified as MRD^pos,, while using LOD and LOQ, we selected the following categories of patients: 1) LOD^neg if RLCc was below LOD (20x100/total no. of events); 2) LOD^pos-LOQ^neg if RLCc was between LOD and LOQ; and 3) LOQ^pos if RLCc was above LOQ (50x100/total no. of events).

Results: The ELN target of 500’000 events was reached in 182/261 (69.6%) patients. Overall, using the predefined AML1310 protocol MRD threshold, 154 (59%) and 107 (41%) were MRD^neg and MRD^pos, respectively, whereas 74 (28.4%), 43 (16.5%) and 144 (54.4%) patients were classified as LOD^neg, LOD^pos-LOQ^neg and LOQ^pos, respectively. Two-year overall survival (OS) was 75.4% vs. 79.8% vs. 66.4% for LOD^neg, LOD^pos-LOQ^neg and LOD^pos, respectively (p=0.1197), and 74.5% vs. 66.4% according to AML1310 protocol 0.035% threshold for MRD^neg and MRD^pos patients, respectively (p=0.3521). Due to superimposable outcome, LOD-LOQ^neg and LOD^pos-LOQ^neg categories were combined. Accordingly, LOD^neg/LOD^pos-LOQ^neg and LOQ^pos groups clearly differed in terms of OS (77% vs. 66.4%, p=0.0437) [FIGURE 1A]. Such a figure was challenged in multivariate analysis (p=0.048, HR 0.628, 95% CI 0.396-0.997) that confirmed the independent role of LOD-LOQ approach in influencing OS.

To enhance the predictivity of LOD-LOQ estimate, we then focused on samples acquisition of which passed the 500’000 events, according to ELN guidelines. Among 182/261 (69.7%) cases with > 500’000 MNC events as denominator, LOD^neg/LOD^pos-LOQ^neg and LOQ^pos subgroups were clearly distinct in terms of OS (2-years OS of 83.5% vs. 69.4%, p=0.009). [FIGURE 1B]

Similarly, also when selecting those patients (158/261 [60.5%]) whose acquisition passed 500’000 CD45+ events, LOD^neg/LOD^pos-LOQ^neg and LOQ^pos showed a different behavior with 2-years OS of 86.7% vs. 69.0%, respectively (p=0.004). [FIGURE 1C]

Finally, when considering the interaction of the 3 LOD-LOQ categories with possible post-remissional strategies, LOD^neg/LOD^pos-LOQ^neg patients submitted to autologous stem cell transplant showed the best 2-years OS (88.9%) as compared to all the other categories (allogeneic stem cells transplant and no graft-based treatments) (p=0.026).

Summary/Conclusion: In conclusion, the use of LOD-LOQ method results in a more sensitive detection of MRD that, in turn, translates in a more accurate recognition of patients with different prognosis. Actually, such an approach allowed to dissect even further the category of patients called MRD^neg according to the AML1310 protocol definition, since MRD^neg subjects who belonged to a “true negative” LOD-LOQ sub-group [LOD^neg/LOD^pos-LOQ^neg] had a better outcome than the other MRD^neg ones. This MRD approach could serve as a useful tool to personalize post-remission strategy in intensively treated AML patients, through selection of high-quality remission patients who may benefit from less intensive post-consolidation therapies.