- Email: [email protected]
Prognostic significance of flow cytometric minimal residual disease assessment after the first induction course in Chinese childhood acute myeloid leukemia
Leukemia Research 37 (2013) 134–138
Contents lists available at SciVerse ScienceDirect
Leukemia Research journal homepage: www.elsevier.com/locate/leukres
Prognostic significance of flow cytometric minimal residual disease assessment after the first induction course in Chinese childhood acute myeloid leukemia Xiao-Jun Xu a,1 , Jian-Hua Feng a,1 , Yong-Min Tang a,b,∗ , Hong-Qiang Shen a,b , Hua Song a , Shi-Long Yang a , Shu-Wen Shi a , Wei-Qun Xu a a b
Division of Hematology–Oncology, Children’s Hospital of Zhejiang University School of Medicine, Hangzhou, PR China Key Laboratory of Reproductive Genetics (Zhejiang University), Ministry of Education, Hangzhou, PR China
a r t i c l e
i n f o
Article history: Received 20 June 2012 Received in revised form 1 November 2012 Accepted 3 November 2012 Available online 28 November 2012 Keywords: Acute myeloid leukemia Childhood Minimal residual disease Relapse Multiparameter flow cytometry
a b s t r a c t Flow cytometry based minimal residual disease (MRD) was evaluated for outcome prediction in childhood acute myeloid leukemia (AML). The median levels of MRD in relapsed and nonrelapsed patients were different after the first induction (0.64% vs. 0.18%, P = 0.030). A cutoff level of ≥0.25% after the first course of induction was correlated with a high risk of relapse in both univariate analysis (5-year cumulative incidence of relapse: 66.8% vs. 21.2%, P = 0.002) and multivariate analyses (hazard ratio: 3.70, 95% CI, 1.23–11.08, P = 0.020). Our results showed that MRD level after the first induction therapy provides important information for risk assessment in childhood AML. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction Acute myeloid leukemia (AML) accounts for about 20% of childhood acute leukemias. With the application of treatment intensification and the improvement in supportive care, the outcome for children with AML has been improved considerably during the past decades [1–4]. However, approximately 30%–40% of patients will eventually relapse even if they have displayed a favorable morphological marrow response, indicating that more accurate and powerful approaches are needed to identify the patients with refractory disease. It is now generally accepted that minimal residual disease (MRD) is a powerful prognostic factor in predicting relapse, refining risk groups for childhood acute lymphoblastic leukemia (ALL). Patients with negative MRD during or after induction reached a 5-year relapse free survival (RFS) of higher than 90% [5,6]. However, the MRD monitoring in AML is not so widely used and seems to be suboptimal. The threshold levels of stratifying patients with different risks and the time points for MRD evaluation have not
∗ Corresponding author at: Division of Hematology–Oncology, Children’s Hospital, Zhejiang University School of Medicine, #57 Zhuganxiang Road, Yan-an Street, Hangzhou 310003, PR China. Tel.: +86 0571 87061007x32460; fax: +86 0571 87033296. E-mail address: Y M [email protected] (Y.-M. Tang). 1 These authors contributed equally to this work. 0145-2126/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.leukres.2012.11.002
been consistent yet [7–9]. Although patients with low MRD level after induction showed a favorable outcome when compared to those with high MRD level, their relapse rate is still very high when compared with ALL [7–9]. On the other hand, the two main techniques for MRD determination, PCR and multiparameter flow cytometry (MFC), showed a coherency of 80% or higher in ALL while they were only approximately 60% in AML [10–12]. These results indicate that the methodology of MRD measurement in AML requires more intensive expertise and should be further developed. Furthermore, more clinical data are needed to optimize the MRD evaluation system in AML. However, up till now, there are only a few reports on the role of MRD in childhood AML and most of them came from western countries [7–9,13,14]. The application and status of MRD monitoring in pediatric AML in resource and expertise limited countries including China are not clear. To the best of our knowledge, this is the first report on MRD detection and its prognostic significance for pediatric AML in Chinese population with such long-term followup. In this prospective study, we used MFC to detect the level of MRD during the follow-up in Chinese pediatric AML patients. Our objective was to determine the prognostic significance of MRD assessment with four-color FCM on clinical outcome in Chinese pediatric AML, and to improve the stratification models used in AML by defining therapy-dependent prognostic parameters from the current study.
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
135
Fig. 1. A representative case showing the gating strategy in AML-MRD measurement. (a) Bone marrow from patients with ALL in complete remission after induction was stained with the antibody combination of CD45/CD19/CD34/CD33 and used as normal control, (b) in the original leukemia, the leukemic cells co-expressed CD33 and CD19. When analyzing the leukemia associated immunophentype (LAIP) in patients with AML, blasts with LAIP of CD45+CD19+CD34+CD33+ are identified as MRD due to their falling in the gating area of abnormality.
2. Patients and methods 2.1. Patients From June 2002 to September 2008, patients fulfilling the following criteria were enrolled in this study: (1) diagnosed as AML except acute promyelocytic leukemia (APL), (2) aged less than 16 years old and (3) intent to treatment for at least 6 months. Patients who could not achieve complete remission (CR) were excluded. Patients abandoned within 6 months after diagnosis were excluded and patients abandoned after 6 months were censored at the date of abandonment. Approval for this study was obtained from the Medical Ethics Committee of our hospital. Informed consent for study participation was obtained from parents or legal guardian of each patient. The diagnosis of AML was based on morphology and immunolophenotyping while cytogenetics and fusion gene examination were also included after 2005. All children were treated according to the modified NPCLC-AML97 protocol as previously published [15]. Subtypes were established based on bone marrow cell morphology and cytochemical stains, including periodic acid-Schiff, peroxidase, chloroacetate and ␣-naphthyl acetate esterases, according to the French–American–British (FAB) classification, and confirmed by immunological methods using MFC. Cytogenetic data were classified according to the Southwest Oncology Group/Eastern Cooperative Oncology Group criteria [16].
2.2. Immunophenotypic studies and detection of MRD Bone marrow samples were sterilely aspirated and placed in test tubes containing preservative-free heparin and mononuclear cells (MNC) were separated after the samples were layered onto the Ficoll-Hypaque medium. The mononuclear cell suspensions at a concentration of 107 cells/ml were prepared in phosphate buffered saline (PBS). The MNC was pre-incubated with human AB serum for 30 min at 37 ◦ C, and was labeled with antibodies individually conjugated with fluorochromes including fluorescein isothiocyanate (FITC), phycoerythrin (PE), peridinin chlorophyll protein (PerCP) and allophycocyanin (APC) (Becton Dickinson, San Jose, CA). After incubation for 20 min, residual erythrocytes were lysed with FACS lysing solution (Becton Dickinson, San Jose, CA, USA). Then, the specimens were washed twice with 2 ml PBS to remove excess antibodies and lysed RBCs. Four-color MFC (FACSCalibur with CellQuest software, Becton Dickinson) was used for acquisition and analysis of data, with application of live gates on the fluorescence/sideward scatter (SSC) (gate 1) and fluorescence/fluorescence plots (gates 2 and 3) (Fig. 1). A minimum of 100,000 events per tube were analyzed. A cluster of at least 10–20 events with the leukemia-associated immunophenotypes (LAIP) were required for MRD
definition. Thus, in the cases of 100,000 bone marrow mononuclear cells (BMMC) were recorded, the sensitivity of MRD detection was at least 1 cell per 10,000 BMMC. Additionally, sensitivity analysis was performed by serial dilutional experiments of leukemic cells seeded into normal BM cells, which confirmed that this technique allowed the detection of at least one leukemic cell among 10,000 normal BMMC (10−4 ). To differentiate normal myeloid progenitors/precursors from leukemic cells, the normal postchemotherapy myeloid hematopoiesis was assessed by using BM samples collected from patients with ALL in complete morphological remission after induction as normal control. The sample processing and staining procedures were similar to AML MRD detection. Background staining levels for all antibody combinations were established using these BM samples. Dual parameter dot plots of fluorescence/SSC or fluorescence/fluorescence were displayed and normal cell populations were defined as normal template. At diagnosis, all samples were screened using four-color technique to confirm the presence of leukemic blasts with aberrant immunophenotypes. After evaluation of the diagnostic results, the patient-tailored labeling was designed and applied to the follow-up samples. The quadruple labeling antibody combinations for MRD detection used in this study were: CD45/CD117/CD34/CD33(25.5%), CD45/CD19/CD34/CD33(12.7%), CD45/CD56/CD34/CD33(16.4%), CD15/CD45/CD34/ CD117(3.6%), CD45/CD117/CD34/CD35(3.6%), CD15/CD117/CD34/CD33(1.8%), CD71/CD117/CD45/CD33(1.8%), CD7/CD117/CD34/CD33(1.8%), CD7/CD117/CD45/ CD117/CD56/CD34/CD33(1.8%), CD45/CD117/CD34/CD35(1.8%), CD33(1.8%), CD11b/CD117/CD45/CD33(1.8%), CD45/CD56/CD15/CD33(1.8%), CD45/CD56/CD14/ CD33(1.8%), CD7/CD56/CD45/CD33(1.8%), CD45/CD56/CD34/CD7(1.8%), HLA-DR/ CD15/CD19/CD34/CD33(5.5%), CD45/CD19/CD34/ CD11b/CD15/CD14(1.8%), CD10(1.8%), HLA-DR/CD15/CD34/CD33(3.6%), CD45/CD11b/CD34/CD33(3.6%) and CD61/CD41/CD34/CD33(1.8%). The conjugation of these antibodies with FITC, PE, PerCP and APC, respectively, were in the order of notation. All the antibody combinations were tested with normal/ALL BM samples and the positive events identified were shown to be 0.01% or below and not clustered. All samples were analyzed with at least two antibody combinations to avoid false negative results in case of antigen shift occurred.
2.3. Statistical analysis For descriptive statistics, median and range, or percentage of cases was calculated. Continuous and categorical data were compared between groups by using the Mann–Whitney U-test and chi-square test, respectively. RFS was measured by the time from achievement of CR to last follow-up or first relapse and censored at
136
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
Fig. 2. Follow-up of patients. CT, chemotherapy.
the first event (death, secondary malignancies) except relapse. Receiver operating characteristic (ROC) curves were derived from the MRD levels and relapse status. In a ROC curve, the sensitivity and specificity were calculated by combining the optimal cut-off value and relapse outcome. Probabilities of survival or relapse data were estimated according to the Kaplan–Meier method, and comparison between the curves was performed using the log-rank test. The Cox proportional hazards regression model on hazard of relapse was applied to study the prognostic value of MRD and other clinical and laboratory features. Hazard ratios with 95% confidence interval and P values were calculated in the models. A P-value < 0.05 (two tailed) was considered to be of statistical significance. Statistical analyses were carried out using Stata ver. 12 software (College Station, TX, USA).
3. Results 3.1. Patients’ characteristics A total of 174 patients were diagnosed as AML from June 2002 to September 2008 in our hospital, with 3 M0, 9 M1, 70 M2, 23 M4, 57 M5 6 M6 and 6 M7. Of the 174 patients, 86 refused any chemotherapy after diagnosis, 4 died of intracranial hemorrhage or infection before chemotherapy, 7 patients did not achieve CR after a maximum of 6 induction courses, 8 abandoned within 6 months after diagnosis and 14 did not receive MRD monitoring. Thus 55 patients were eligible for MRD evaluation (Fig. 2). The characteristics of the 55 patients are given in Table 1. The clinical features including gender, age, WBC count, blast count, FAB classification and cytogenetics were all comparable between the non-eligible patients and the MRD determination cohort (data not shown). The median follow-up time was 5.1 (range: 3.1–8.4) years. Twentythree (41.8%) relapsed ranging from 2.0 month after CR to 3.8 years with a median of 1.0 years, and 28 (50.9%) remained in continuous CR (CCR). The 5-year RFS for the whole cohort was 56.7 ± 6.9%. Patients relapsed and those remained in CCR were compared for their clinical characteristics (age, sex, WBC, hemoglobin, platelets, percentage of peripheral blast and cytogenetics). No significant differences were observed between those two groups in terms of almost all parameters tested except the blast level (Table 1). During the study period, 162 samples were collected at the 4 time points (TP), including 46 after induction chemotherapy (TP1), 42 after second course of chemotherapy (TP2), 40 at week 12 (TP3) and 34 on month 6 (TP4), with an average of three samples per patient. 3.2. MRD levels at various time points The MRD distributions of all the four TPs were shown in Fig. 3. The proportions of patients with negative MRD (<0.01%) increased from 10.9% at TP1 to 23.5% at TP4, while the percentages of patients
Fig. 3. MRD levels during follow up. MRD levels after induction chemotherapy (TP1), after second course of chemotherapy (TP2), at week 12 (TP3) and in month 6 (TP4).
with very high MRD level (≥1.0%) decreased from 26.1% at TP1 to 8.8% at TP4. The estimated 5-year cumulative incidence of relapse (CIR) showed a trend of increase with increasing MRD levels at TP1: 20.0 ± 17.9%, 25.9 ± 12.9%, 47.1 ± 12.1% and 64.8 ± 15.2% in groups with MRD < 0.01%, 0.01%–0.1%, 0.1%–1.0% and ≥1.0% (P = 0.197). At other TPs, the CIRs were comparable among groups of MRD <0.01%, 0.01%–0.1%, and 0.1%–1.0%, while only patients with MRD ≥1.0% presented much higher relapse rate (all >60%) when compared with other MRD level groups (data not shown). The median level of MRD detected in AML patients by MFC was 0.24% at TP1, 0.22% at TP2, 0.22% at TP3 and 0.18% at TP4. Besides, we compared MRD levels at all the time points in CCR group and in relapse group. MRD level was much higher in relapsed patients than that in CCR patients at TP1 (0.64% vs. 0.18%, P = 0.030), while the MRD levels between the two groups at other TPs were comparable (TP2, 0.22% vs. 0.18%, P = 0.741; TP3, 0.23% vs. 0.22%, P = 0.345, TP4, 0.28% vs. 0.16%, P = 0.701). 3.3. Prognostic significance of MRD assessment after the first induction therapy Using a cut-off value of 0.1% for MFC MRD analysis, we found a trend of differences between patients with MRD <0.1% and MRD ≥ 0.1% in the 5-year RFS at TP1 (75.6 ± 10.6% vs. 45.9 ± 9.6%, P = 0.052) while no differences were found at other TPs (data not shown). We then performed a ROC analysis to determine the optimal threshold that could split AML patients into two groups with different MRD levels and correlated with relapse status. The AUC for
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
137
Table 1 Patient characteristics. Characteristics
All patients (n = 55)
Relapsed (n = 23)
CCR (n = 32)
P-value
Sex (m/f) Age (years), median (range) WBC (×109 L−1 ), median (range) Hemoglobin (gL−1 ), median (range) Platelets (×109 L–1 ), median (range) % Peripheral blast, median (range) Cytogeneticsa Favorable Intermediate Unfavorable FAB classification (%) M0 M1 M2 M4 M5 M6 M7
34/21 7.9 (1.5–16.1) 16.7 (1.7–163.6) 73.0 (32.0–120.0) 44.0 (3.9–197.0) 8.0 (0.0–85.0)
16/7 8.4 (2.8–16.1) 25.2 (1.7–154.0) 65.0 (32.0–120.0) 29.0 (12.0–197.0) 17.0 (0–85.0)
18/14 7.5 (1.5–13.7) 11.2 (2.3–163.6) 76.5 (48.0–113.0) 48.5 (3.9–167.0) 5.0 (0–79.0)
0.316 0.352 0.061 0.168 0.137 0.039
a b
15 22 6 1 (1.8) 3 (5.5) 28 (50.9) 7 (12.7) 13 (23.6) 2 (3.6) 1 (1.8)
8 6 4 0 (0.0) 3 (9.4) 16 (50.0) 4 (12.5) 7 (21.9) 1 (3.1) 1 (3.1)
7 16 2
b
0.264
1 (4.3) 0 (0.0) 12 (52.2) 3 (13.0) 6 (26.1) 1 (4.3) 0 (0.0)
Available for 43 of 55 patients. Comparison between the two cytogenetic subgroups: favorable versus unfavorable and intermediate.
4. Discussion
Fig. 4. Cumulative incidences of relapse (CIR) according to MRD after the first induction therapy. Patients with MRD equal to or higher than 0.25% showed much higher CIR when compared to those with MRD less than 0.25% (66.8 ± 10.5% vs. 21.2 ± 8.4%, P = 0.002).
the four TPs were 0.690 (95% confidence interval [CI], 0.523–0.857, P = 0.030) at TP1, 0.530 (95% CI, 0.346–0.714, P = 0.741) at TP2; 0.591 (95% CI, 0.401–0.781, P = 0.334) at TP3 and 0.540 (95% CI, 0.323–0.757, P = 0.697) at TP4, indicating that only MRD at TP1 were predictive of relapse. The optimal cut-off value for TP1 was 0.27%, yielding sensitivity and specificity for predicting relapse of 73.7% (95% CI, 51.2–88.2) and 70.4% (95% CI, 51.5–84.2), respectively. Thus, we use the MRD level of 0.25% as a threshold value to discriminate patients with different risk of relapse at TP1. Univariate analysis using the Kaplan–Meier method for prognostic risk assessment revealed that MRD at TP1 (≥0.25% vs. <0.25%) had a significant impact on relapse, with 5-year CIR of 66.8 ± 10.5% and 21.2 ± 8.4% (P = 0.002), respectively (Fig. 4). Of the other disease characteristics, i.e. cytogenetic characteristics, WBC count, blast in peripheral blood and age, WBC count (>15 × 109 L−1 or not) and blast in peripheral blood (>15%) showed significant influence on 5year CIR in univariate analysis (WBC: 60.1 ± 9.7% and 26.3 ± 8.5%, P = 0.023, blast: 67.0 ± 11.9% and 30.3 ± 8.0%, P = 0.028). However, only the MRD level at TP1 was the independent prognostic factor in a multivariate Cox analysis including the above parameters, with a hazard ratio of 3.70 (95% CI, 1.23–11.08, P = 0.020) for MRD level equal to or higher than 0.25% of BMMC when compared to those with MRD less than 0.25%.
MRD detection with MFC in clinical remission has been proposed as a feasible strategy to stratify the risk of relapse after induction and it has also been correlated with clinical outcome for patients with AML in some studies [7–9,13,14,17,18]. Data from our study suggest that the high levels of MRD (≥0.25%) after the first induction was associated with a significant likelihood of subsequent relapse and a short duration RFS. Thus, the magnitude of cytoreduction after the first induction appears to be critical to the clinical outcome of disease. In addition, we tried to define a threshold for each treatment point. With the exception of the cut-off value of 0.25% at TP1, we were unable to find differences of RFS between patients with different MRD levels in late phases of CR (TP2, TP3 and TP4). In this study, the BM specimens were processed using Ficoll gradients to isolate MNCs for flow cytometry analysis. However, most clinical flow cytometry labs now perform staining on whole bone marrow or blood specimens. Although our method is more complicated and may cause a possible loss of blast, it can minimize the disturbance from other cells such as mature neutrophils and macrophages, which can non-specifically bind the antibody via their Fc receptors even after BSA blocking and some even showed auto-fluorescence. Furthermore, the leukemia cells can be enriched after ficoll centrifugation which allows us to detect the leukemia cells more easily in the samples. The results in this study indicated that the MRD measurement using this method was powerful and feasible to identify the patients with different risks of relapse. However, due to the different methodology we used, our result cannot be used to directly compare with those from other labs using whole marrow staining. Theoretically, this method may lead to a relatively higher threshold for prognostic significance of AML MRD. To the best of our knowledge, this is the first AML-MRD study conducted in Chinese pediatric population in mainland of China. Although the race, medical resources, treatment protocol and technology in our cohort were different from those of the western countries, our results were similar to theirs: the level of MRD after the first course of chemotherapy is significantly related to clinical outcome [8,9]. Although the clinical outcome in this study was inferior to those of the western countries, the MRD distribution at different time points was similar with other study [9]. In contrast to ALL patients, the majority of whom can achieve negative MRD during the entire course of chemotherapy, only a small fraction (∼25%) of patients with AML can achieve negative MRD (<0.01%).
138
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
Due to the high residual disease level in most patients, MRD < 0.01% is rarely used as an optimal cut-off value for risk stratification in AML patients and the level of MRD that is most predictive of outcome in childhood AML is inconsistent, from 0.1% to 0.5% [7–9]. In our study, although cut-off values of 0.1% and 0.5% tended to discriminate patients with different relapse risks (5-year CIR for MRD ≥0.5% and MRD < 0.5% were 66.9 ± 11.4% and 26.4 ± 8.6%, P = 0.005), a threshold of 0.25% allowed for MRD detection with optimal sensitivity and specificity in this cohort of patients. This difference may be explained by the different therapeutic regimens used and divergent technical approaches employed such as utilizing patient-tailored labeling and sample processing. As we have discussed in our previous reports, abandonment is a big obstacle to the improvement of the cure rate of leukemia in China [15]. In this study, it is the leading cause of treatment failure, which accounted approximately half of the patients. Financial difficulties and lack of confidence that this disease can be cured are the main issues contributed to abandonment. In this circumstance, if the patients are proved to be with high probability of curable disease by his/her low MRD level at a very early stage of chemotherapy, they may be more confident and adhere to treatment. From this perspective, the MRD determination after first induction course might benefit the patients with low risk disease by reducing the abandonment rate. In conclusion, we found that flow cytometry MRD assessment after the first course of induction chemotherapy is a significant prognostic factor in Chinese childhood AML, which can be used to discriminate patients with superior outcome from those with higher risk of relapse. Employing risk group stratification might enable patients with a lower risk of relapse to avoid exposure to unnecessary therapy and improve clinical outcome of patients with high MRD levels by more intensive therapeutic intervention. Therefore, the level of MRD after the first induction course may be implemented into therapeutic strategies in childhood AML. Conflict of interest statement The authors declare no conflict of interest. Acknowledgements The authors would like to thank Mrs. Bai-Qin Qian at the Hematology–Oncology Laboratory in the Children’s Hospital of Zhejiang University School of Medicine for her excellent technical support. This study was supported in part by grants from the National Natural Science Foundation of China (Nos. 30971283, 31100638, 81170502), the Science and Technology Bureau of Zhejiang Province (2007C23007), the Zhejiang Provincial Natural Science Foundation of China (Y2110020), the Health Bureau of Zhejiang Province (2007B122) and the Ph.D. Programs Foundation of Ministry of Education of China (20110101120138).
Contributions. Y.-M.T. was the principal investigator and takes the primary responsibility for the paper. X.-J.X. and J.-H.F. collected and analyzed the data, wrote the manuscript draft which was amended by all the authors. H.S., S.-L.Y., S.-W.S., W.-Q.X. collected the MRD sample. H.-Q.S. performed the MRD measurement with flow cytometry. All the authors approved final version submitted. References [1] Meshinchi S, Arceci RJ. Prognostic factors and risk-based therapy in pediatric acute myeloid leukemia. Oncologist 2007;12:341–55. [2] Rubnitz JE. Childhood acute myeloid leukemia. Curr Treat Options Oncol 2008;9:95–105. [3] Kaspers GJ, Zwaan CM. Pediatric acute myeloid leukemia: towards high-quality cure of all patients. Haematologica 2007;92:1519–32. [4] Pession A, Rondelli R, Basso G, Rizzari C, Testi AM, Fagioli F, et al. Treatment and long-term results in children with acute myeloid leukaemia treated according to the AIEOP AML protocols. Leukemia 2005;19:2043–53. [5] Borowitz MJ, Devidas M, Hunger SP, Bowman WP, Carroll AJ, Carroll WL, et al. Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia and its relationship to other prognostic factors: a children’s oncology group study. Blood 2008;111:5477–85. [6] Xu XJ, Tang YM, Shen HQ, Song H, Yang SL, Shi SW, et al. Day 22 of induction therapy is important for minimal residual disease assessment by flow cytometry in childhood acute lymphoblastic leukemia. Leuk Res 2012;36:1022–7. [7] Sievers EL, Lange BJ, Alonzo TA, Gerbing RB, Bernstein ID, Smith FO, et al. Immunophenotypic evidence of leukemia after induction therapy predicts relapse: results from a prospective Children’s Cancer Group study of 252 patients with acute myeloid leukemia. Blood 2003;101:3398–406. [8] Coustan-Smith E, Ribeiro RC, Rubnitz JE, Razzouk BI, Pui CH, Pounds S, et al. Clinical significance of residual disease during treatment in childhood acute myeloid leukaemia. Br J Haematol 2003;123:243–52. [9] van der Velden VH, van der Sluijs-Geling A, Gibson BE, te Marvelde JG, Hoogeveen PG, Hop WC, et al. Clinical significance of flowcytometric minimal residual disease detection in pediatric acute myeloid leukemia patients treated according to the DCOG ANLL97/MRC AML12 protocol. Leukemia 2010;24:1599–606. [10] Gaipa G, Cazzaniga G, Valsecchi MG, Panzer-Grumayer R, Buldini B, Silvestri D, et al. Time point-dependent concordance of flow cytometry and RQ-PCR inminimal residual disease detection in childhood acute lymphoblasticleukemia. Haematologica 2012;97:1582–93. [11] Neale GA, Coustan-Smith E, Stow P, Pan Q, Chen X, Pui CH, et al. Comparative analysis of flow cytometry and polymerase chain reaction for the detection of minimal residual disease in childhood acute lymphoblastic leukemia. Leukemia 2004;18:934–8. [12] Kern W, Schoch C, Haferlach T, Voskova D, Hiddemann W, Schnittger S. Complemental roles for multiparameter flow cytomety and quantitative RT-PCR for the quantification of minimal residual disease in patients with acute myeloid leukemia. Blood 2003;102:65a–70a. [13] Sievers EL, Lange BJ, Buckley JD, Smith FO, Wells DA, Daigneault-Creech CA, et al. Prediction of relapse of pediatric acute myeloid leukemia by use of multidimensional flow cytometry. J Natl Cancer Inst 1996;88:1483–8. [14] Langebrake C, Creutzig U, Dworzak M, Hrusak O, Mejstrikova E, Griesinger F, et al. Residual disease monitoring in childhood acute myeloid leukemia by multiparameter flow cytometry: the MRD-AML-BFM study group. J Clin Oncol 2006;24:3686–92. [15] Xu XJ, Tang YM, Song H, Yang SL, Shi SW, Wei J. Long-term outcome of childhood acute myeloid leukemia in a developing country: experience from a children’s hospital in China. Leuk Lymphoma 2010;51:2262–9. [16] Mrozek K, Heerema NA, Bloomfield CD. Cytogenetics in acute leukemia. Blood Rev 2004;18:115–36. [17] San Miguel JF, Vidriales MB, Lopez-Berges C, Diaz-Mediavilla J, Gutierrez N, Canizo C, et al. Early immunophenotypical evaluation of minimal residual disease in acute myeloid leukemia identifies different patient risk groups and may contribute to postinduction treatment stratification. Blood 2001;98:1746–51. [18] Feller N, van der Pol MA, van Stijn A, Weijers GW, Westra AH, Evertse BW, et al. MRD parameters using immunophenotypic detection methods are highly reliable in predicting survival in acute myeloid leukaemia. Leukemia 2004;18:1380–90.
Contents lists available at SciVerse ScienceDirect
Leukemia Research journal homepage: www.elsevier.com/locate/leukres
Prognostic significance of flow cytometric minimal residual disease assessment after the first induction course in Chinese childhood acute myeloid leukemia Xiao-Jun Xu a,1 , Jian-Hua Feng a,1 , Yong-Min Tang a,b,∗ , Hong-Qiang Shen a,b , Hua Song a , Shi-Long Yang a , Shu-Wen Shi a , Wei-Qun Xu a a b
Division of Hematology–Oncology, Children’s Hospital of Zhejiang University School of Medicine, Hangzhou, PR China Key Laboratory of Reproductive Genetics (Zhejiang University), Ministry of Education, Hangzhou, PR China
a r t i c l e
i n f o
Article history: Received 20 June 2012 Received in revised form 1 November 2012 Accepted 3 November 2012 Available online 28 November 2012 Keywords: Acute myeloid leukemia Childhood Minimal residual disease Relapse Multiparameter flow cytometry
a b s t r a c t Flow cytometry based minimal residual disease (MRD) was evaluated for outcome prediction in childhood acute myeloid leukemia (AML). The median levels of MRD in relapsed and nonrelapsed patients were different after the first induction (0.64% vs. 0.18%, P = 0.030). A cutoff level of ≥0.25% after the first course of induction was correlated with a high risk of relapse in both univariate analysis (5-year cumulative incidence of relapse: 66.8% vs. 21.2%, P = 0.002) and multivariate analyses (hazard ratio: 3.70, 95% CI, 1.23–11.08, P = 0.020). Our results showed that MRD level after the first induction therapy provides important information for risk assessment in childhood AML. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction Acute myeloid leukemia (AML) accounts for about 20% of childhood acute leukemias. With the application of treatment intensification and the improvement in supportive care, the outcome for children with AML has been improved considerably during the past decades [1–4]. However, approximately 30%–40% of patients will eventually relapse even if they have displayed a favorable morphological marrow response, indicating that more accurate and powerful approaches are needed to identify the patients with refractory disease. It is now generally accepted that minimal residual disease (MRD) is a powerful prognostic factor in predicting relapse, refining risk groups for childhood acute lymphoblastic leukemia (ALL). Patients with negative MRD during or after induction reached a 5-year relapse free survival (RFS) of higher than 90% [5,6]. However, the MRD monitoring in AML is not so widely used and seems to be suboptimal. The threshold levels of stratifying patients with different risks and the time points for MRD evaluation have not
∗ Corresponding author at: Division of Hematology–Oncology, Children’s Hospital, Zhejiang University School of Medicine, #57 Zhuganxiang Road, Yan-an Street, Hangzhou 310003, PR China. Tel.: +86 0571 87061007x32460; fax: +86 0571 87033296. E-mail address: Y M [email protected] (Y.-M. Tang). 1 These authors contributed equally to this work. 0145-2126/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.leukres.2012.11.002
been consistent yet [7–9]. Although patients with low MRD level after induction showed a favorable outcome when compared to those with high MRD level, their relapse rate is still very high when compared with ALL [7–9]. On the other hand, the two main techniques for MRD determination, PCR and multiparameter flow cytometry (MFC), showed a coherency of 80% or higher in ALL while they were only approximately 60% in AML [10–12]. These results indicate that the methodology of MRD measurement in AML requires more intensive expertise and should be further developed. Furthermore, more clinical data are needed to optimize the MRD evaluation system in AML. However, up till now, there are only a few reports on the role of MRD in childhood AML and most of them came from western countries [7–9,13,14]. The application and status of MRD monitoring in pediatric AML in resource and expertise limited countries including China are not clear. To the best of our knowledge, this is the first report on MRD detection and its prognostic significance for pediatric AML in Chinese population with such long-term followup. In this prospective study, we used MFC to detect the level of MRD during the follow-up in Chinese pediatric AML patients. Our objective was to determine the prognostic significance of MRD assessment with four-color FCM on clinical outcome in Chinese pediatric AML, and to improve the stratification models used in AML by defining therapy-dependent prognostic parameters from the current study.
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
135
Fig. 1. A representative case showing the gating strategy in AML-MRD measurement. (a) Bone marrow from patients with ALL in complete remission after induction was stained with the antibody combination of CD45/CD19/CD34/CD33 and used as normal control, (b) in the original leukemia, the leukemic cells co-expressed CD33 and CD19. When analyzing the leukemia associated immunophentype (LAIP) in patients with AML, blasts with LAIP of CD45+CD19+CD34+CD33+ are identified as MRD due to their falling in the gating area of abnormality.
2. Patients and methods 2.1. Patients From June 2002 to September 2008, patients fulfilling the following criteria were enrolled in this study: (1) diagnosed as AML except acute promyelocytic leukemia (APL), (2) aged less than 16 years old and (3) intent to treatment for at least 6 months. Patients who could not achieve complete remission (CR) were excluded. Patients abandoned within 6 months after diagnosis were excluded and patients abandoned after 6 months were censored at the date of abandonment. Approval for this study was obtained from the Medical Ethics Committee of our hospital. Informed consent for study participation was obtained from parents or legal guardian of each patient. The diagnosis of AML was based on morphology and immunolophenotyping while cytogenetics and fusion gene examination were also included after 2005. All children were treated according to the modified NPCLC-AML97 protocol as previously published [15]. Subtypes were established based on bone marrow cell morphology and cytochemical stains, including periodic acid-Schiff, peroxidase, chloroacetate and ␣-naphthyl acetate esterases, according to the French–American–British (FAB) classification, and confirmed by immunological methods using MFC. Cytogenetic data were classified according to the Southwest Oncology Group/Eastern Cooperative Oncology Group criteria [16].
2.2. Immunophenotypic studies and detection of MRD Bone marrow samples were sterilely aspirated and placed in test tubes containing preservative-free heparin and mononuclear cells (MNC) were separated after the samples were layered onto the Ficoll-Hypaque medium. The mononuclear cell suspensions at a concentration of 107 cells/ml were prepared in phosphate buffered saline (PBS). The MNC was pre-incubated with human AB serum for 30 min at 37 ◦ C, and was labeled with antibodies individually conjugated with fluorochromes including fluorescein isothiocyanate (FITC), phycoerythrin (PE), peridinin chlorophyll protein (PerCP) and allophycocyanin (APC) (Becton Dickinson, San Jose, CA). After incubation for 20 min, residual erythrocytes were lysed with FACS lysing solution (Becton Dickinson, San Jose, CA, USA). Then, the specimens were washed twice with 2 ml PBS to remove excess antibodies and lysed RBCs. Four-color MFC (FACSCalibur with CellQuest software, Becton Dickinson) was used for acquisition and analysis of data, with application of live gates on the fluorescence/sideward scatter (SSC) (gate 1) and fluorescence/fluorescence plots (gates 2 and 3) (Fig. 1). A minimum of 100,000 events per tube were analyzed. A cluster of at least 10–20 events with the leukemia-associated immunophenotypes (LAIP) were required for MRD
definition. Thus, in the cases of 100,000 bone marrow mononuclear cells (BMMC) were recorded, the sensitivity of MRD detection was at least 1 cell per 10,000 BMMC. Additionally, sensitivity analysis was performed by serial dilutional experiments of leukemic cells seeded into normal BM cells, which confirmed that this technique allowed the detection of at least one leukemic cell among 10,000 normal BMMC (10−4 ). To differentiate normal myeloid progenitors/precursors from leukemic cells, the normal postchemotherapy myeloid hematopoiesis was assessed by using BM samples collected from patients with ALL in complete morphological remission after induction as normal control. The sample processing and staining procedures were similar to AML MRD detection. Background staining levels for all antibody combinations were established using these BM samples. Dual parameter dot plots of fluorescence/SSC or fluorescence/fluorescence were displayed and normal cell populations were defined as normal template. At diagnosis, all samples were screened using four-color technique to confirm the presence of leukemic blasts with aberrant immunophenotypes. After evaluation of the diagnostic results, the patient-tailored labeling was designed and applied to the follow-up samples. The quadruple labeling antibody combinations for MRD detection used in this study were: CD45/CD117/CD34/CD33(25.5%), CD45/CD19/CD34/CD33(12.7%), CD45/CD56/CD34/CD33(16.4%), CD15/CD45/CD34/ CD117(3.6%), CD45/CD117/CD34/CD35(3.6%), CD15/CD117/CD34/CD33(1.8%), CD71/CD117/CD45/CD33(1.8%), CD7/CD117/CD34/CD33(1.8%), CD7/CD117/CD45/ CD117/CD56/CD34/CD33(1.8%), CD45/CD117/CD34/CD35(1.8%), CD33(1.8%), CD11b/CD117/CD45/CD33(1.8%), CD45/CD56/CD15/CD33(1.8%), CD45/CD56/CD14/ CD33(1.8%), CD7/CD56/CD45/CD33(1.8%), CD45/CD56/CD34/CD7(1.8%), HLA-DR/ CD15/CD19/CD34/CD33(5.5%), CD45/CD19/CD34/ CD11b/CD15/CD14(1.8%), CD10(1.8%), HLA-DR/CD15/CD34/CD33(3.6%), CD45/CD11b/CD34/CD33(3.6%) and CD61/CD41/CD34/CD33(1.8%). The conjugation of these antibodies with FITC, PE, PerCP and APC, respectively, were in the order of notation. All the antibody combinations were tested with normal/ALL BM samples and the positive events identified were shown to be 0.01% or below and not clustered. All samples were analyzed with at least two antibody combinations to avoid false negative results in case of antigen shift occurred.
2.3. Statistical analysis For descriptive statistics, median and range, or percentage of cases was calculated. Continuous and categorical data were compared between groups by using the Mann–Whitney U-test and chi-square test, respectively. RFS was measured by the time from achievement of CR to last follow-up or first relapse and censored at
136
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
Fig. 2. Follow-up of patients. CT, chemotherapy.
the first event (death, secondary malignancies) except relapse. Receiver operating characteristic (ROC) curves were derived from the MRD levels and relapse status. In a ROC curve, the sensitivity and specificity were calculated by combining the optimal cut-off value and relapse outcome. Probabilities of survival or relapse data were estimated according to the Kaplan–Meier method, and comparison between the curves was performed using the log-rank test. The Cox proportional hazards regression model on hazard of relapse was applied to study the prognostic value of MRD and other clinical and laboratory features. Hazard ratios with 95% confidence interval and P values were calculated in the models. A P-value < 0.05 (two tailed) was considered to be of statistical significance. Statistical analyses were carried out using Stata ver. 12 software (College Station, TX, USA).
3. Results 3.1. Patients’ characteristics A total of 174 patients were diagnosed as AML from June 2002 to September 2008 in our hospital, with 3 M0, 9 M1, 70 M2, 23 M4, 57 M5 6 M6 and 6 M7. Of the 174 patients, 86 refused any chemotherapy after diagnosis, 4 died of intracranial hemorrhage or infection before chemotherapy, 7 patients did not achieve CR after a maximum of 6 induction courses, 8 abandoned within 6 months after diagnosis and 14 did not receive MRD monitoring. Thus 55 patients were eligible for MRD evaluation (Fig. 2). The characteristics of the 55 patients are given in Table 1. The clinical features including gender, age, WBC count, blast count, FAB classification and cytogenetics were all comparable between the non-eligible patients and the MRD determination cohort (data not shown). The median follow-up time was 5.1 (range: 3.1–8.4) years. Twentythree (41.8%) relapsed ranging from 2.0 month after CR to 3.8 years with a median of 1.0 years, and 28 (50.9%) remained in continuous CR (CCR). The 5-year RFS for the whole cohort was 56.7 ± 6.9%. Patients relapsed and those remained in CCR were compared for their clinical characteristics (age, sex, WBC, hemoglobin, platelets, percentage of peripheral blast and cytogenetics). No significant differences were observed between those two groups in terms of almost all parameters tested except the blast level (Table 1). During the study period, 162 samples were collected at the 4 time points (TP), including 46 after induction chemotherapy (TP1), 42 after second course of chemotherapy (TP2), 40 at week 12 (TP3) and 34 on month 6 (TP4), with an average of three samples per patient. 3.2. MRD levels at various time points The MRD distributions of all the four TPs were shown in Fig. 3. The proportions of patients with negative MRD (<0.01%) increased from 10.9% at TP1 to 23.5% at TP4, while the percentages of patients
Fig. 3. MRD levels during follow up. MRD levels after induction chemotherapy (TP1), after second course of chemotherapy (TP2), at week 12 (TP3) and in month 6 (TP4).
with very high MRD level (≥1.0%) decreased from 26.1% at TP1 to 8.8% at TP4. The estimated 5-year cumulative incidence of relapse (CIR) showed a trend of increase with increasing MRD levels at TP1: 20.0 ± 17.9%, 25.9 ± 12.9%, 47.1 ± 12.1% and 64.8 ± 15.2% in groups with MRD < 0.01%, 0.01%–0.1%, 0.1%–1.0% and ≥1.0% (P = 0.197). At other TPs, the CIRs were comparable among groups of MRD <0.01%, 0.01%–0.1%, and 0.1%–1.0%, while only patients with MRD ≥1.0% presented much higher relapse rate (all >60%) when compared with other MRD level groups (data not shown). The median level of MRD detected in AML patients by MFC was 0.24% at TP1, 0.22% at TP2, 0.22% at TP3 and 0.18% at TP4. Besides, we compared MRD levels at all the time points in CCR group and in relapse group. MRD level was much higher in relapsed patients than that in CCR patients at TP1 (0.64% vs. 0.18%, P = 0.030), while the MRD levels between the two groups at other TPs were comparable (TP2, 0.22% vs. 0.18%, P = 0.741; TP3, 0.23% vs. 0.22%, P = 0.345, TP4, 0.28% vs. 0.16%, P = 0.701). 3.3. Prognostic significance of MRD assessment after the first induction therapy Using a cut-off value of 0.1% for MFC MRD analysis, we found a trend of differences between patients with MRD <0.1% and MRD ≥ 0.1% in the 5-year RFS at TP1 (75.6 ± 10.6% vs. 45.9 ± 9.6%, P = 0.052) while no differences were found at other TPs (data not shown). We then performed a ROC analysis to determine the optimal threshold that could split AML patients into two groups with different MRD levels and correlated with relapse status. The AUC for
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
137
Table 1 Patient characteristics. Characteristics
All patients (n = 55)
Relapsed (n = 23)
CCR (n = 32)
P-value
Sex (m/f) Age (years), median (range) WBC (×109 L−1 ), median (range) Hemoglobin (gL−1 ), median (range) Platelets (×109 L–1 ), median (range) % Peripheral blast, median (range) Cytogeneticsa Favorable Intermediate Unfavorable FAB classification (%) M0 M1 M2 M4 M5 M6 M7
34/21 7.9 (1.5–16.1) 16.7 (1.7–163.6) 73.0 (32.0–120.0) 44.0 (3.9–197.0) 8.0 (0.0–85.0)
16/7 8.4 (2.8–16.1) 25.2 (1.7–154.0) 65.0 (32.0–120.0) 29.0 (12.0–197.0) 17.0 (0–85.0)
18/14 7.5 (1.5–13.7) 11.2 (2.3–163.6) 76.5 (48.0–113.0) 48.5 (3.9–167.0) 5.0 (0–79.0)
0.316 0.352 0.061 0.168 0.137 0.039
a b
15 22 6 1 (1.8) 3 (5.5) 28 (50.9) 7 (12.7) 13 (23.6) 2 (3.6) 1 (1.8)
8 6 4 0 (0.0) 3 (9.4) 16 (50.0) 4 (12.5) 7 (21.9) 1 (3.1) 1 (3.1)
7 16 2
b
0.264
1 (4.3) 0 (0.0) 12 (52.2) 3 (13.0) 6 (26.1) 1 (4.3) 0 (0.0)
Available for 43 of 55 patients. Comparison between the two cytogenetic subgroups: favorable versus unfavorable and intermediate.
4. Discussion
Fig. 4. Cumulative incidences of relapse (CIR) according to MRD after the first induction therapy. Patients with MRD equal to or higher than 0.25% showed much higher CIR when compared to those with MRD less than 0.25% (66.8 ± 10.5% vs. 21.2 ± 8.4%, P = 0.002).
the four TPs were 0.690 (95% confidence interval [CI], 0.523–0.857, P = 0.030) at TP1, 0.530 (95% CI, 0.346–0.714, P = 0.741) at TP2; 0.591 (95% CI, 0.401–0.781, P = 0.334) at TP3 and 0.540 (95% CI, 0.323–0.757, P = 0.697) at TP4, indicating that only MRD at TP1 were predictive of relapse. The optimal cut-off value for TP1 was 0.27%, yielding sensitivity and specificity for predicting relapse of 73.7% (95% CI, 51.2–88.2) and 70.4% (95% CI, 51.5–84.2), respectively. Thus, we use the MRD level of 0.25% as a threshold value to discriminate patients with different risk of relapse at TP1. Univariate analysis using the Kaplan–Meier method for prognostic risk assessment revealed that MRD at TP1 (≥0.25% vs. <0.25%) had a significant impact on relapse, with 5-year CIR of 66.8 ± 10.5% and 21.2 ± 8.4% (P = 0.002), respectively (Fig. 4). Of the other disease characteristics, i.e. cytogenetic characteristics, WBC count, blast in peripheral blood and age, WBC count (>15 × 109 L−1 or not) and blast in peripheral blood (>15%) showed significant influence on 5year CIR in univariate analysis (WBC: 60.1 ± 9.7% and 26.3 ± 8.5%, P = 0.023, blast: 67.0 ± 11.9% and 30.3 ± 8.0%, P = 0.028). However, only the MRD level at TP1 was the independent prognostic factor in a multivariate Cox analysis including the above parameters, with a hazard ratio of 3.70 (95% CI, 1.23–11.08, P = 0.020) for MRD level equal to or higher than 0.25% of BMMC when compared to those with MRD less than 0.25%.
MRD detection with MFC in clinical remission has been proposed as a feasible strategy to stratify the risk of relapse after induction and it has also been correlated with clinical outcome for patients with AML in some studies [7–9,13,14,17,18]. Data from our study suggest that the high levels of MRD (≥0.25%) after the first induction was associated with a significant likelihood of subsequent relapse and a short duration RFS. Thus, the magnitude of cytoreduction after the first induction appears to be critical to the clinical outcome of disease. In addition, we tried to define a threshold for each treatment point. With the exception of the cut-off value of 0.25% at TP1, we were unable to find differences of RFS between patients with different MRD levels in late phases of CR (TP2, TP3 and TP4). In this study, the BM specimens were processed using Ficoll gradients to isolate MNCs for flow cytometry analysis. However, most clinical flow cytometry labs now perform staining on whole bone marrow or blood specimens. Although our method is more complicated and may cause a possible loss of blast, it can minimize the disturbance from other cells such as mature neutrophils and macrophages, which can non-specifically bind the antibody via their Fc receptors even after BSA blocking and some even showed auto-fluorescence. Furthermore, the leukemia cells can be enriched after ficoll centrifugation which allows us to detect the leukemia cells more easily in the samples. The results in this study indicated that the MRD measurement using this method was powerful and feasible to identify the patients with different risks of relapse. However, due to the different methodology we used, our result cannot be used to directly compare with those from other labs using whole marrow staining. Theoretically, this method may lead to a relatively higher threshold for prognostic significance of AML MRD. To the best of our knowledge, this is the first AML-MRD study conducted in Chinese pediatric population in mainland of China. Although the race, medical resources, treatment protocol and technology in our cohort were different from those of the western countries, our results were similar to theirs: the level of MRD after the first course of chemotherapy is significantly related to clinical outcome [8,9]. Although the clinical outcome in this study was inferior to those of the western countries, the MRD distribution at different time points was similar with other study [9]. In contrast to ALL patients, the majority of whom can achieve negative MRD during the entire course of chemotherapy, only a small fraction (∼25%) of patients with AML can achieve negative MRD (<0.01%).
138
X.-J. Xu et al. / Leukemia Research 37 (2013) 134–138
Due to the high residual disease level in most patients, MRD < 0.01% is rarely used as an optimal cut-off value for risk stratification in AML patients and the level of MRD that is most predictive of outcome in childhood AML is inconsistent, from 0.1% to 0.5% [7–9]. In our study, although cut-off values of 0.1% and 0.5% tended to discriminate patients with different relapse risks (5-year CIR for MRD ≥0.5% and MRD < 0.5% were 66.9 ± 11.4% and 26.4 ± 8.6%, P = 0.005), a threshold of 0.25% allowed for MRD detection with optimal sensitivity and specificity in this cohort of patients. This difference may be explained by the different therapeutic regimens used and divergent technical approaches employed such as utilizing patient-tailored labeling and sample processing. As we have discussed in our previous reports, abandonment is a big obstacle to the improvement of the cure rate of leukemia in China [15]. In this study, it is the leading cause of treatment failure, which accounted approximately half of the patients. Financial difficulties and lack of confidence that this disease can be cured are the main issues contributed to abandonment. In this circumstance, if the patients are proved to be with high probability of curable disease by his/her low MRD level at a very early stage of chemotherapy, they may be more confident and adhere to treatment. From this perspective, the MRD determination after first induction course might benefit the patients with low risk disease by reducing the abandonment rate. In conclusion, we found that flow cytometry MRD assessment after the first course of induction chemotherapy is a significant prognostic factor in Chinese childhood AML, which can be used to discriminate patients with superior outcome from those with higher risk of relapse. Employing risk group stratification might enable patients with a lower risk of relapse to avoid exposure to unnecessary therapy and improve clinical outcome of patients with high MRD levels by more intensive therapeutic intervention. Therefore, the level of MRD after the first induction course may be implemented into therapeutic strategies in childhood AML. Conflict of interest statement The authors declare no conflict of interest. Acknowledgements The authors would like to thank Mrs. Bai-Qin Qian at the Hematology–Oncology Laboratory in the Children’s Hospital of Zhejiang University School of Medicine for her excellent technical support. This study was supported in part by grants from the National Natural Science Foundation of China (Nos. 30971283, 31100638, 81170502), the Science and Technology Bureau of Zhejiang Province (2007C23007), the Zhejiang Provincial Natural Science Foundation of China (Y2110020), the Health Bureau of Zhejiang Province (2007B122) and the Ph.D. Programs Foundation of Ministry of Education of China (20110101120138).
Contributions. Y.-M.T. was the principal investigator and takes the primary responsibility for the paper. X.-J.X. and J.-H.F. collected and analyzed the data, wrote the manuscript draft which was amended by all the authors. H.S., S.-L.Y., S.-W.S., W.-Q.X. collected the MRD sample. H.-Q.S. performed the MRD measurement with flow cytometry. All the authors approved final version submitted. References [1] Meshinchi S, Arceci RJ. Prognostic factors and risk-based therapy in pediatric acute myeloid leukemia. Oncologist 2007;12:341–55. [2] Rubnitz JE. Childhood acute myeloid leukemia. Curr Treat Options Oncol 2008;9:95–105. [3] Kaspers GJ, Zwaan CM. Pediatric acute myeloid leukemia: towards high-quality cure of all patients. Haematologica 2007;92:1519–32. [4] Pession A, Rondelli R, Basso G, Rizzari C, Testi AM, Fagioli F, et al. Treatment and long-term results in children with acute myeloid leukaemia treated according to the AIEOP AML protocols. Leukemia 2005;19:2043–53. [5] Borowitz MJ, Devidas M, Hunger SP, Bowman WP, Carroll AJ, Carroll WL, et al. Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia and its relationship to other prognostic factors: a children’s oncology group study. Blood 2008;111:5477–85. [6] Xu XJ, Tang YM, Shen HQ, Song H, Yang SL, Shi SW, et al. Day 22 of induction therapy is important for minimal residual disease assessment by flow cytometry in childhood acute lymphoblastic leukemia. Leuk Res 2012;36:1022–7. [7] Sievers EL, Lange BJ, Alonzo TA, Gerbing RB, Bernstein ID, Smith FO, et al. Immunophenotypic evidence of leukemia after induction therapy predicts relapse: results from a prospective Children’s Cancer Group study of 252 patients with acute myeloid leukemia. Blood 2003;101:3398–406. [8] Coustan-Smith E, Ribeiro RC, Rubnitz JE, Razzouk BI, Pui CH, Pounds S, et al. Clinical significance of residual disease during treatment in childhood acute myeloid leukaemia. Br J Haematol 2003;123:243–52. [9] van der Velden VH, van der Sluijs-Geling A, Gibson BE, te Marvelde JG, Hoogeveen PG, Hop WC, et al. Clinical significance of flowcytometric minimal residual disease detection in pediatric acute myeloid leukemia patients treated according to the DCOG ANLL97/MRC AML12 protocol. Leukemia 2010;24:1599–606. [10] Gaipa G, Cazzaniga G, Valsecchi MG, Panzer-Grumayer R, Buldini B, Silvestri D, et al. Time point-dependent concordance of flow cytometry and RQ-PCR inminimal residual disease detection in childhood acute lymphoblasticleukemia. Haematologica 2012;97:1582–93. [11] Neale GA, Coustan-Smith E, Stow P, Pan Q, Chen X, Pui CH, et al. Comparative analysis of flow cytometry and polymerase chain reaction for the detection of minimal residual disease in childhood acute lymphoblastic leukemia. Leukemia 2004;18:934–8. [12] Kern W, Schoch C, Haferlach T, Voskova D, Hiddemann W, Schnittger S. Complemental roles for multiparameter flow cytomety and quantitative RT-PCR for the quantification of minimal residual disease in patients with acute myeloid leukemia. Blood 2003;102:65a–70a. [13] Sievers EL, Lange BJ, Buckley JD, Smith FO, Wells DA, Daigneault-Creech CA, et al. Prediction of relapse of pediatric acute myeloid leukemia by use of multidimensional flow cytometry. J Natl Cancer Inst 1996;88:1483–8. [14] Langebrake C, Creutzig U, Dworzak M, Hrusak O, Mejstrikova E, Griesinger F, et al. Residual disease monitoring in childhood acute myeloid leukemia by multiparameter flow cytometry: the MRD-AML-BFM study group. J Clin Oncol 2006;24:3686–92. [15] Xu XJ, Tang YM, Song H, Yang SL, Shi SW, Wei J. Long-term outcome of childhood acute myeloid leukemia in a developing country: experience from a children’s hospital in China. Leuk Lymphoma 2010;51:2262–9. [16] Mrozek K, Heerema NA, Bloomfield CD. Cytogenetics in acute leukemia. Blood Rev 2004;18:115–36. [17] San Miguel JF, Vidriales MB, Lopez-Berges C, Diaz-Mediavilla J, Gutierrez N, Canizo C, et al. Early immunophenotypical evaluation of minimal residual disease in acute myeloid leukemia identifies different patient risk groups and may contribute to postinduction treatment stratification. Blood 2001;98:1746–51. [18] Feller N, van der Pol MA, van Stijn A, Weijers GW, Westra AH, Evertse BW, et al. MRD parameters using immunophenotypic detection methods are highly reliable in predicting survival in acute myeloid leukaemia. Leukemia 2004;18:1380–90.