Prognostic Value of Minimal Residual Disease Measured By Flow Cytometry in Two Cohorts of Infants with Acute Lymphoblastic Leukemia Treated with MLL-Baby and Interfant Protocols in Large Multicenter Networks

Acute lymphoblastic leukemia (ALL) in infants is a relatively rare disease with peculiar biological features, high frequency of KMT2A gene rearrangements and grim prognosis. Even with new therapeutic approaches, event-free survival (EFS) in infants with ALL does not exceed 50%. Currently large cooperative studies of infant ALL have been promoted by the Interfant and MLL-Baby networks. Minimal residual disease (MRD) monitoring is considered a strong tool for optimizing management of childhood ALL. In contrast to older children in this age group the prognostic impact of MRD detected by multicolor flow cytometry (MFC) is still unclear. Aim of the present study was to evaluate the prognostic value of MFC MRD measurement during induction in infants with ALL treated with Interfant-99 and Interfant-06 protocols in AIEOP centers in Italy as well as MLL-Baby protocol in Russia and Belarus.

Patients and methods. Two independent groups of patients were investigated: study cohort of 139 consecutive infants with newly diagnosed ALL enrolled between September 2003 and April 2016 in Russian and Belorussian centers with MLL-Baby protocol and validation group of 146 ones enrolled in AIEOP centers in Interfant trials during the same period. By availability of MFC MRD data obtained at day 15 and/or end of induction (EOI), 81 and 86 patients of study and validation cohorts respectively were selected for outcome evaluation. In 61 MLL-Baby patients (75.3%) and 61 AIEOP cases (70.9%) different types of KMT2A gene rearrangements were identified. All patients were diagnosed as BCP-ALL, except one with cortical T-ALL. Overall, day 15 samples were studied in 64 MLL-Baby patients and 73 AIEOP cases while EOI samples in 75 and 63 cases respectively. MRD detection was performed in Reference Laboratories in Ekaterinburg, Minsk and Padua according the BFM AIEOP FLOW Network SOP. MRD negativity was defined as <0.01% of all bone marrow nucleated cells.

Results. Patients were stratified according to the AIEOP-BFM-ALL day 15 stratification usually used for older children into three risk groups: standard risk (SR: MRD<0.1%), intermediate risk (IR: MRD 0.1% to 10%) and high risk (HR: MRD≥10%). Patients' distribution was similar in both study and validation cohorts: 34.4% and 32.9% in SR, 53.1% and 54.5% in ImR, 12.5% and 12.3% in HR respectively. At EOI significant differences in MRD-positive and MRD-negative patients' distribution was observed in different protocols: 44 (58.7%) and 31 (41.3%) cases respectively for MLL-Baby, but 17 (27.0%) and 46 (73.0%) cases respectively for AIEOP group. It was observed that KMT2A-rearranged cases in both trials have slower MRD response compared to the children with wild type KMT2A. In study cohort the 22 SR patients had a 5-year EFS and cumulative incidence of relapse (CIR) significantly better than other ones, thus we considered two major groups of patients with different outcome: SR with 5-year EFS 67.4%, standard error (SE) 10.2; CIR 23.3%, SE 9.4 and non-SR with 5-year EFS 30.8%, SE 7.2; CIR 52.6%, SE 7.9, (p=0.0039 and p=0.0229, respectively). Difference between these groups was observed also in KMT2A-rearranged cases (n=49) both for 5-year EFS (60.0%, SE 12.7 and 23.2%, SE 7.3, p=0.0160) and in 5-year CIR (33.3%, SE 12.7 and 59.2%, SE 8.8, p=0.0881). Analysis of outcome in validation cohort confirmed these data. In study cohort outcome of children being MRD-negative at EOI (n=31, 5-year EFS 60.8%, SE 8.8 and CIR 29.3%, SE 8.4) was significantly better than that of MRD-positive patients (n=44, 5-year EFS 31.1%, SE 7.1 and CIR 57.6%, SE 7.8 with p=0.0153 and p=0.0267, respectively). Outcome by EOI MFC MRD in the validation cohort is generally in keeping with that of the study cohort. Interestingly, in AIEOP cohort MFC data showed a prognostic impact also in KMT2A-rearranged subgroup. In multivariate analysis with KMT2A-status, each MRD time-point data showed independent impact on the risk of relapse.

Conclusion. Our data was obtained by well-harmonized MFC MRD monitoring in a large group of infants with ALL treated in a multicenter setting with two different protocols. In spite of differences in therapy, we observed strong and independent prognostic impact of MFC MRD both at day 15 and at EOI regardless the protocol applied. We can conclude that MFC MRD can be used in combination with KMT2A-status to improve treatment allocation in future protocols.