A total of 390 pediatric patients admitted and initially diagnosed as B-ALL in Jining First People's Hospital from August 2010 to May 2016 were selected, and they were diagnosed via bone marrow morphology, immunology, cytogenetics and molecular biology (MICM) classification. All pediatric patients were followed up; this study followed the ethical standards set by the Institutional Review Board and was approved by The Ethics Committee of Jining First People's Hospital. Written, informed consent was obtained from all patients or their parents prior to their inclusion within the study.

The risk of leukemia was stratified according to the age of pediatric patients in the initial diagnosis, peripheral leukocyte count, abnormalities in cytogenetics and molecular biology, whether they were sensitive to prednisone, bone marrow remission status at 33 days and MRD at 33 days; the therapeutic regimen was based on the CCLG-2008 protocol (modified BFM ALL-95 protocol) (13), and the diagnosis and efficacy were based on ‘Diagnosis and Treatment Recommendations of Pediatric Acute Leukemia (Revised Draft)’.

The risk was stratified according to the age of pediatric patients in the initial diagnosis, peripheral leukocyte count, bone marrow puncture results (immunophenotyping, chromosome and 29 fusion genes) and treatment effects, and ALL pediatric patients could be divided into standard risk (SR) group, intermediate risk (IR) group and high risk (HR) group.

Bone marrow (2–3 ml) was taken from pediatric patients for anticoagulation to separate the karyocytes, followed by immunofluorescence labeling using the living-cell index method (FITC or PE). The antibodies used were all mouse monoclonal antibodies and the dilution was 1/100: CD2 (cat. no. 563820); CD3 (cat. no. 565066); CD5 (cat. no. 564648); CD7 (cat. no. 566119); CD10 (cat. no. 564959); CD19 (cat. no. 564978); CD20 (cat. no. 564918); CD22 (cat. no. 563940); cCD79 (cat. no. 555302); cCD3 (cat. no. 565065); HLA-DR (cat. no. 564516); TdT (cat. no. 565229); MPO (cat. no. 556035) and Smlg (cat. no. 564555). All were purchased from BD Biosciences (San Diego, CA, USA). Flow cytometer (Beckman Coulter, Brea, CA, USA) was used; antigen-expressing cells in juvenile cell group >20% indicated the positive.

The chromosomes were taken from the bone marrow of pediatric patients, and the chromosome samples were prepared using the direct method or 24 h short-term culture. The karyotype was analyzed using the conventional R-banding technique. The chromosome abnormality was explained based on the International System for Human Cytogenetic Nomenclature (ISCN1985) (14). Bone marrow (2–3 ml) with heparin anticoagulation was taken in the initial diagnosis, and the mononuclear cells were extracted using lymphocyte analytic liquid (Ficoll). Twenty-nine fusion genes [including ETV6/RUNXI (TEL/AML1), MLL rearrangement, HOX11, E2A/PBX1, and BCR/ABL] were detected via nested reverse transcription polymerase chain reaction (RT-PCR).

Bone marrow (2–3 ml) with heparin anticoagulation was taken from pediatric patients in the initial diagnosis before chemotherapy, and the karyocytes were separated routinely and counted. Then the karyocytes were stained using CD45-FITC grading, CD10-PE, CD34-Percp and CD19-APC monoclonal antibodies; the antibody combination was effective if the leukemia cells and normal bone marrow cells could be distinguished. At 33 days and 12 weeks after induction chemotherapy, the bone marrow samples were taken from pediatric patients, and the mononuclear cells were separated using Ficoll, followed by four-color fluorescence labeling detection (CD58, CD66c, CD38, CD123 and TdT were added based on CD34/CD19/CD10); the original data in the initial diagnosis and screening were called up; in the two-parameter diagram, if there were cells in the original area of leukemia cells, it indicated the minimal residual leukemia; the proportion of these residual cells in total bone marrow mononuclear cells was the result of monitoring MRD.

Among the newly-diagnosed pediatric patients selected, except those with mature-type B-ALL (n=12) who were treated as non-Hodgkin's lymphoma, all the other pediatric patients were treated according to CCLG-2008-ALL protocol and chemotherapy regimens with different intensities according to different risk degrees as follows: i) remission induction: pretreatment with prednisone for 7 days, VDLD (vincristine, daunorubicin, L-asparaginase or pegaspargase, dexamethasone) regimen in induction; ii) early intensive treatment: CAM (cyclophosphamide, cytarabine, 6-mercaptopurine or thioguanine) program (SR group × one round; IR and HR groups × two rounds); iii) consolidation therapy: SR and IR groups: HD-methotrexate (MTX) + 6-MP program (2.0 g/m² MTX in SR group, 5.0 g/m² MTX in IR group); HR group: HR-1, HR-2 and HR-3 (a total of two rounds); iv) delayed enhancement: the same VDLD and CAM programs as above [two rounds of delayed enhancement for IR group (one round of maintenance chemotherapy between the two rounds of delayed enhancement)]; v) maintenance treatment: 6-MP + MTX. The single or triple chemotherapy drugs were regularly injected intrathecally to prevent the central nervous system leukemia.

Re-induction therapy: VMDP (vincristine, mitoxantrone, dexamethasone, pegaspargase). Intensive consolidation therapy: FLAG (fludarabine, HD-cytarabine, granulocyte-stimulating factor) or CAM + 6-MP. Transplantation or continuous treatment [delayed enhancement (vincristine, dexamethasone, methotrexate, 6-MP, cyclophosphamide, cytarabine, etoposide) + maintenance chemotherapy (vincristine, dexamethasone, methotrexate, 6-MP)]. Extramedullary relapse: local radiotherapy after chemotherapy. CNSL: head radiotherapy 1800 cGy, 150 cGy/day, 12 days; TL: testicular local radiotherapy 2400 cGy, 200 cGy/day, 12 days.

Complete remission (CR): the primitive and juvenile lymphocytes <5% on the bone marrow smear after induction chemotherapy. Relapse: including simple intramedullary relapse, extramedullary relapse (central nervous system or testicular relapse) and intra-extramedullary relapse.

Patients were followed up until December 31, 2015, the median follow-up time was 22.7 months (2.8–62.4 months), and there were a total of 80 cases of relapse. EFS: the time from relapse to relapse again or death. Statistical Product and Service Solutions (SPSS, Chicago, IL, USA) 17.0 Software Package was used for the comparison of 3-year EFS rate of B-ALL pediatric patients in intramedullary relapse group at different time points; Kaplan-Meier survival curve analysis method was used for the comparison of 3-year EFS rate in different risk degrees; bilateral log-rank test was used for the survival situations in different groups. The independent-samples t-test was used for the clinical characteristics between B-ALL relapse group and non-relapse group, and Chi-square test was used for the comparison of constituent ratio of clinical characteristics between the two groups. P<0.05 was considered to indicate a statistically significant difference.

The general data of subjects in the two groups are shown in Table I. There was a significant difference in the proportion of B-ALL pediatric patients with different risk stratification in relapse group and non-relapse group; the pairwise comparison among SR group, IR group and HR group showed that the difference in the proportion of relapse in B-ALL pediatric patients was not statistically significant between SR group and IR group, but the difference was statistically significant compared with that in HR group (P=0.013). The comparison of pediatric patients in B-ALL relapse group and non-relapse group revealed that there was no statistically significant difference in the sex composition between the two groups (P=0.421, P>0.05), but there were statistically significant differences in the age and ratio of white blood cell (WBC) ≥50 × 10⁹/l in the initial diagnosis (P=0.031, P=0.032, P<0.05).

The cytogenetics and molecular biology results of pediatric patients with relapse and non-relapse in B-ALL group are shown in Tables II and III. For chromosome analysis, we did not detect significant difference between two groups while for gene analysis, fusion gene rate in the relapse group was significantly higher than that in the non-relapse group.

MRD of pediatric patients with B-ALL relapse and non-relapse was monitored at 33 days and 12 weeks after chemotherapy, and they were divided according to different MRD levels for analysis and comparison (Tables IV and V). At day 33, there were more children relapsed in the MRD ≥10⁻² group than that in MRD <10⁻² group (P<0.05). At week 12, there were more children relapsed in the MRD ≥10⁻³ group than that in MRD <10⁻³ group (P<0.05). The relapse rate in the <10⁻⁴ group was lower than non-relapse rate (P<0.05). Moreover, the ratio in 10⁻⁴−10⁻³, 10⁻³−10⁻² and ≥10⁻² in relapse group was higher than that in non-relapse group. As a result, MRD <10⁻⁴ could be thought as a cut-off point of low relapse rate.

The relapse site and time of pediatric patients in the relapse group and their relationship with risk typing are shown in Tables VI and VII. As defined by Berlin-Frankfurt-Munster-ALL Group, early relapse: the time from initial diagnosis to relapse was ≤18 months; medium relapse: the time from initial diagnosis to relapse was >18 months and CR <30 months; late relapse: CR >30 months.

The EFS curve of B-ALL pediatric patients with early, medium and late bone marrow relapse, the EFS curve of B-ALL pediatric patients with bone marrow relapse in SR group, IR group and HR group are shown in Figs. 1 and 2. The 3-year event-free survival (EFS) rates of pediatric patients with early, medium and late B-ALL relapse were 12.5±7.8, 33.1±9.8 and 63.6±6.1%, respectively, and the prognosis of late relapse was significantly better than that of early relapse (P<0.001). The 3-year EFS rates of pediatric patients with bone marrow relapse in standard risk group, intermediate risk group and high risk group were 29.1±6.9, 31.3±6.5 and 28.3±6.3%, respectively; there were no statistically significant differences (P=0.387, P>0.05).