Value of Flow Cytometry for MRD-Based Relapse Prediction in B-Cell Precursor Acute Lymphoblastic Leukemia in a Multi-Center Setting

Abstract

Background: PCR of rearranged antigen receptor genes is the method of choice for MRD quantification in ALL. Although FCM-MRD is faster and biologically more informative than PCR, the analysis requires a high level of training. The only larger published studies using FCM-MRD based stratification (Borowitz, Blood, 2008 and 2015) showed a clear association with clinical outcome in BCP-ALL. However, MRD analyses were centralized and these studies included only one MRD-based stratification (MRD levels at the end of induction). Patients and methods: We examined FCM-MRD as stratification tool in BCP-ALL at various timepoints in a large-scale multicenter (18 MRD centers) study. A total of 1487 patients with BCP-ALL (1298 children (younger than 18 years) and 189 adults (18-45 years) are included in the study and were treated according to the NOPHO ALL2008 protocol between July 2008 and February 2016. The median follow-up time for patients in first remission was 51 months (IQR 32-75). MRD was measured by FCM and/or real time quantitative PCR on days 15, 29 (end of induction) and 79 (for standard (SR) and intermediate risk (IR) patients) and prior to and after high risk blocks. A 6-colour FCM analysis including 3 standardized antibody combinations was used and performed in 18 laboratories. Patients were stratified by FCM-MRD, or by PCR-MRD if no FCM-MRD marker was available. End-of-induction MRD (cut-off 10-3) was used to stratify patients to standard risk (SR) vs intermediate risk (IR) or IR vs high risk consolidation therapy (in case of WBC > 100 x 109/L at diagnosis). Patients with MRD >=2.5x10-1 on day 15 were stratified to high risk block therapy. Patients with MRD >=5x10-2 on day 29 or day 79/post high risk-2 block MRD >=10-3 were stratified to HSCT. Primary outcomes were 5year event-free survival (5y EFS) and 5year cumulative incidence of relapse (5y CIR). Results: Only two patients (0.14% of total) had neither an informative FCM nor a PCR marker, and an informative FCM marker combination for MRD monitoring was identified in 96.2% of patients. There was a significant correlation between FCM- and PCR-MRD levels on day 15 (r=0.77, p<0.0001, n=153) and 29 (r=0.81, p<0.0001, n=140). Based on FCM-MRD only, the median MRD level on day 15, 29 and 79/post high risk-2 block was 5x10-3, 1.1x10-4, and below detection limit, respectively. Adults had significantly higher MRD levels at all time-points (p<0.0001 for day 15 and 29, p=0.0019 for day 79, Mann-Whitney). The 5y EFS was 86.1% (95% CI 84.1-88.1) with a 5y CIR of 9.5% (95% CI 7.8-11.3, n=1487). The day 29 FCM-MRD level was closely associated with clinical outcome and a higher hazard of relapse was seen independently for a FCM-MRD >=10-3 (hazard ratio (HR) 2.4, CI 1.6-3.7, p<0.0001), age>18 year (HR 3.0, CI 1.7-5.3, p<0.0001), WBC>=100 (HR 2.7, CI 1.6-4.6, p=0.0001), and B-other (HR 2.1, CI 1.2-3.5, p=0.0052) or high risk B-ALL cytogenetic aberration (rearranged KMT2A/iAMPchr21/hypodiploid) (HR 3.2, CI 1.6-6.1, p=0.0006) (multivariate cause-specific Cox regression, n=1328). Patients with a day 79 FCM-MRD >=10-4 and <10-3 had a significantly higher CIR (22.1%, CI 10.8-33.5%, n=68) compared to FCM-MRD <10-4 (7.5%, CI 2.1-12.8%, n=110) or undetectable (6.3%, CI 4.5-8.2%, n=999, p=0.0087 for FCM-MRD >=10-4 and <10-3vs <10-4 or undetectable). After adjusting for WBC, age, and the day 29 FCM-MRD level, a day 79 FCM-MRD >=10-4 and <10-3 was still significantly associated with a worse 5y CIR for non-transplanted patients (HR 2.3, CI 1.19-4.36, p=0.012 compared to undetectable FCM-MRD, n=1171). Patients with day 15 FCM-MRD <10-3 had a significantly better 5y EFS (92.0%, CI 89.2-95.0%) and CIR (3.9%, CI 1.7-6.1%, n=432) than patients with FCM-MRD >=10-3 and <2.5x10-1, who had a 5y EFS of 85.5% (CI 82.7-88.3%, p=0.0016, n=837) and a 3-fold higher 5y CIR (11.0%, CI 8.4-13.5%, p<0.0001, n=432). Among patients with day 15 FCM-MRD<10-3, the relapse incidence was comparable for patients with FCM-MRD 10-4 - <10-3 and below 10-4 (CIR 3.6, CI 0.5-6.7 vs. CIR 4.1, CI 1.0-7.2, p=0.83, n=432). Conclusion: FCM-MRD performed in a multi-center setting is a clinically useful method for disease monitoring and MRD-based treatment stratification in BCP-ALL. Moreover, FCM-MRD is a reliable indicator of outcome in BCP-ALL independently of other key risk factors. Residual disease >=10-4 and <10-3 at day 79 in SR/IR patients not allocated to HSCT further identifies patients with a high risk of relapse. Disclosures No relevant conflicts of interest to declare.