Abstract
The epidemiology of infant acute lymphoblastic leukemia (ALL), hypodiploid ALL, and mixed-phenotype acute leukemia (MPAL) in Canada is unknown. The main objective was to describe the prevalence, prognostic factors, and outcomes of three rare and high-risk ALL subtypes in Canada. This is a retrospective study using the Cancer in Young People-Canada (CYP-C) database. Event-free survival (EFS) and overall survival (OS) were described by the Kaplan–Meier method and compared using the log-rank test. Among 2626 children aged 0–14 years diagnosed with B-cell acute lymphoblastic leukemia (B-ALL) between 2001 and 2018, 227 (8.6%) patients were identified to be infant ALL (n = 139), hypodiploid ALL (n = 43), or MPAL (n = 45). The 5-year EFS/OS was significantly worse in the infant ALL subgroup compared to that of hypodiploid ALL and MPAL. For the entire cohort, presenting White blood cells (WBCs) ≥50 × 109/L was independently associated with worse EFS/OS.
Introduction
Acute lymphoblastic leukemia (ALL) represents the most common childhood cancer and undoubtedly portrays a true success story in the field of pediatric oncology. The overall survival rates of children with ALL went from less than 10% in the 1960s and now exceed 90% with current chemotherapy regimens [Citation1]. This significant outcome improvement stems from the optimization of multi-agent chemotherapy regimens via randomized clinical trials, better risk stratification as a result of an improved understanding of leukemia biology, and careful monitoring of minimal residual disease (MRD) [Citation2]. Despite the significant survival improvement, relapse still occurs in 15–20% of patients and outcomes remain dismal following relapse, rendering ALL one of the leading causes of cancer-related death among children and young adults [Citation2,Citation3]. Some ALL subtypes remain at high risk of treatment failure despite therapy intensification, such as infant ALL, hypodiploid ALL, and mixed-phenotype acute leukemia (MPAL). These three ALL subtypes are rare, each representing less than 5% of childhood ALL in large international cohorts [Citation4–6]. The prevalence and outcomes of infant ALL, hypodiploid ALL, and MPAL are unknown in Canada. The primary objective of this study aims to describe the prevalence, clinical features, and outcomes of infant ALL, hypodiploid ALL, and MPAL in Canada using the Cancer in Young People-Canada (CYP-C) national registry.
Design/method
Study population
We included all children aged less than 15 years at leukemia diagnosis, diagnosed with infant ALL, hypodiploid ALL, and MPAL between 1 January 2001 and 1 January 2018, treated at one of the 17 pediatric oncology centers in Canada, and registered in CYP-C. Children diagnosed with infant ALL were identified using the International Classification of Diseases for Oncology (ICD-O) codes for ALL (9805-9, 9811-19, and 9836) and age at diagnosis <1 year. Patients’ diagnoses were determined based on the treating physician’s working diagnosis, and no central revision occurred. Children with hypodiploid ALL and MPAL were identified using the following ICD-O codes (9816/3 for hypodiploid, 9805/3, and 9808/3 for MPAL). Hypodiploid ALL was also identified and cross-validated with the ICD-O code in cases with the hypodiploidy classification or 44 chromosomes or fewer in the cytogenetics’ description. Patients were also identified and/or verified with MPAL’s ICD-O codes when MPAL-related classification (i.e. biphenotypic, bilineage, mixed phenotype, ambiguous lineage) was added to their disease description.
Data source
We used the CYP-C database, which is a population-based registry that includes all children (<15 years old) with cancers diagnosed and treated at 1 of the 17 tertiary pediatric oncology centers in Canada since 2001 [Citation7,Citation8]. CYP-C is a collaboration between the federal Public Health Agency of Canada, the C17 Council, and the Canadian Partnership against Cancer. Data are collected and submitted to CYP-C in 2 ways. For the 5 Ontario centers, data are transferred to CYP-C via the Pediatric Oncology Group of Ontario Networked Information System, which is a provincial population-based registry that predates CYP-C. The 12 centers outside of Ontario enter data directly into CYP-C. Elements captured by both databases include demographic variables (sex, date of birth, postal code, and race), diagnostic details, time to diagnosis and treatment, treatment plan details, and outcomes such as relapse, second malignancy, and death. The CYP-C program achieves high-quality data through the supervision of each site’s data manager, monthly review teleconferences, and annual face-to-face training combined with site audits.
Covariates definition
National Cancer Institute (NCI) risk was based on age at diagnosis and presenting white blood cell (WBC) count. Patients were classified as NCI standard-risk (SR) if age at diagnosis was >1 year or <10 years or presented white blood cell (WBC) count < 50,000 cells/L, the remainder was classified as high-risk (HR). Central nervous system (CNS) disease was defined by the presence of lymphoblasts and ≥5 leukocytes/µL in the diagnostic lumbar puncture (CNS-3), cranial nerve palsies, or identification of leukemic infiltrates of the leptomeninges or brain parenchyma. Favorable cytogenetics included cases with the ETV6-RUNX1 fusion, high hyperdiploidy defined as >50 chromosomes, double trisomies of chromosomes 4 and 10, or triple trisomies of chromosomes 4, 10, and 17; unfavorable cytogenetics encompassed cases with BCR-ABL1, KMT2A rearrangement or hypodiploidy defined as <44 chromosomes. The exact modal number of chromosomes at leukemia diagnosis was not available. The presence of intrachromosomal amplification of chromosome 21 (iAMP21) was considered unfavorable cytogenetics for patients diagnosed after 2018. Hematopoietic stem cell transplantation (HSCT) in first complete remission was assumed when the date of HSCT was prior to the date of the first relapse. For MPAL patients, acute myeloid leukemia (AML)-directed induction was defined by treatment initiation according to AML-based induction chemotherapy followed by alternative treatment, whereas AML-directed overall treatment was considered when treatment components were predominantly based on an AML protocol, irrespective of whether treatment was initiated with either AML or ALL-directed therapy.
Statistical analysis
Categorical and continuous patient characteristics were presented for the entire cohort and by subtype (infant, hypodiploid, or MPAL). We performed an analysis of event-free survival (EFS) and overall survival (OS) for all patients (no distinction between groups) by the Kaplan–Meier method. EFS was defined from the time of initial ALL diagnosis to the first event (relapse, secondary malignancy, death from any cause) or last contact for those who were event-free. OS was defined from the time of diagnosis to death from any cause or date of the last follow-up for those who were alive. We also performed an analysis of EFS and OS for each subtype by the Kaplan–Meier method and compared the curves using the log-rank test. For multiple comparisons, we used the Sidak adjustment method to calculate the pairwise comparison results of log-rank tests. We conducted univariate and multivariable Cox proportional hazard regression models of EFS and OS for all patients and then by subtype. The predictors of outcome tested in univariate analyses were selected according to the existing literature for each leukemia subtype. It was determined a priori that factors associated with the outcome of interest with a p-value <0.2 in univariate analysis were included in the list of predictors for the multivariable Cox proportional hazard model.
Results
Among 2626 children <15 years of age with a diagnosis of B-lineage ALL from 2001 to 2018 in Canada, 227 (8.6%) patients were diagnosed with one of the three ALL subtypes of interest: 139 (5.3%) infant ALL, 43 (1.6%) hypodiploid ALL and 45 (1.7%) MPAL.
Patients’ characteristics
The clinical characteristics of the entire cohort are summarized in . The median age and (WBC) count at diagnosis were 7 [4–10] months and 105.5 [23.8–358.1] × 109/L for infant ALL; 7 [3–14] years and 12.8 [5.0–26.1] × 109/L for hypodiploid ALL; 7 [4–12] years and 12.9 [3.8–54.4] × 109/L for MPAL. Infant ALL had the highest presenting WBC at diagnosis compared to the other 2 groups (p < 0.0001). More than half of hypodiploid ALL (55.8%) and MPAL (53.3%) patients were high-risk (HR) according to the NCI Rome criteria. The male-to-female ratio was evenly distributed within each subtype. Caucasian ethnicity predominated within all subtypes, with 47.5% in infant ALL, 69.8% in hypodiploid ALL, and 55.6% in MPAL. The majority of patients were CNS-1 at diagnosis; however, CNS-3 status was highest among infant ALL at 15.8% (p = 0.001). Approximately a third of patients were registered for a clinical trial during treatment (n = 76, 33.5%), and about half of them were treated before the year 2010 (n = 109, 48.0%). Patients with hypodiploid ALL were more likely to have been enrolled in a clinical trial (53.5%; p < 0.0001). More than a third of patients (n = 83, 36.6%) underwent (HSCT) at some point in their treatment. 42 (18.5%) patients of the entire cohort went to HSCT prior to first relapse: 16 (11.5%) infant ALL, 11 (25.6%) hypodiploid ALL, and 15 (33.3%) MPAL.
Table 1. Baseline patient characteristics.
Leukemia outcomes
For our entire study population, the 5-year EFS and OS were 47.4 ± 3.4% and 65.2 ± 3.4%, respectively (). The 5-year EFS was significantly worse in the infant ALL subgroup (38.9 ± 4.6%) compared to that of hypodiploid ALL (63.0 ± 8.6%; p = 0.001) and MPAL (58.8 ± 8.1%; p = 0.003). The 5-year OS was significantly inferior in infant ALL (58.9 ± 4.6%) compared to hypodiploid ALL (75.2 ± 7.2%; p = 0.027) and MPAL (74.9 ± 7.0%; p = 0.034).
Figure 1. (A) Event-free survival and (B) overall survival of the entire cohort. (C) Event-free survival and (D) overall survival by subtype.
Predictors of event-free and overall survival
We evaluated different predictors of EFS and OS for each of the three ALL subtypes and for the entire cohort, respectively ( and ). In the infant ALL subgroup, age <3 months, WBC ≥100 × 109/L at diagnosis, and CNS-3 status were significant in the univariate analysis for EFS and OS, but only age <3 months (HR: 4.01, 95% CI: [1.75–9.15]; p = 0.001 and HR: 4.51, 95% CI: [1.81–11.27]; p = 0.001) and presenting WBC ≥100 × 109/L (HR: 3.85, 95% CI: [1.68–8.83]; p = 0.002 and HR: 3.39, 95% CI: [1.34–8.56]; p = 0.010) remained statistically significant in the multivariable analyses for both EFS and OS, respectively. Additionally, KMT2A rearrangement was a significant adverse predictor in the univariate analysis of OS in the infant ALL subgroup (HR: 7.12, 95% CI: [1.67–30.31]; p = 0.008), but lost statistical significance in the multivariable analysis (p = 0.051, data not shown). In the hypodiploid ALL subgroup, there were no significant prognostic factors in the univariate and multivariable analyses for both EFS and OS. For the MPAL subgroup, age ≥10 years, enrollment in a clinical trial, treatment prior to the year 2010 and overall treatment following AML-directed protocols were significant in the univariate analyses for EFS and OS. In the multivariable analysis, overall treatment following AML protocols (HR: 29.10, 95% CI: [3.38–250.39]; p = 0.002 and HR: 22.88, 95% CI: [2.53–207.10]; p = 0.005) represented an independent adverse predictor of EFS and OS, respectively. For the entire study population, WBC ≥50 × 109/L and CNS-3 at diagnosis were significant in univariate analyses of EFS and OS; however, only WBC ≥50 × 109/L was independently associated with worse EFS (HR: 2.40, 95% CI: [1.58–3.63]; p < 0.001) and OS (HR: 2.48, 95% CI: [1.27–4.82]; p = 0.008) in the multivariable analysis. CNS-3 at diagnosis conferred an unfavorable prognostic factor in the multivariable analysis of OS (HR: 2.58, 95% CI: [0.58–6.21]; p = 0.034) but not EFS (HR: 1.58, 95% CI: [0.86–2.90]; p = 0.142).
Table 2. Univariate and multivariable hazard models for event-free survival in infant acute lymphoblastic leukemia (ALL), hypodiploid ALL, mixed-phenotype acute leukemia (MPAL) and the entire cohort.
Table 3. Univariate and multivariable hazard models of overall survival in infant acute lymphoblastic leukemia (ALL), hypodiploid ALL, mixed-phenotype acute leukemia (MPAL) and the entire cohort.
Discussion
This is the first population-based study to describe the prevalence, prognostic factors, and outcomes of three rare and unfavorable ALL subtypes in Canada. Overall, infant ALL, hypodiploid ALL, and MPAL constitute 8.6% of children <15 years diagnosed with B-cell acute lymphoblastic leukemia (B-ALL) in Canada between 2001 and 2018. The 5-year EFS and OS for the entire cohort were 47.4% and 65.2%, respectively. Within a similar time period and using the CYP-C database, Strahlendorf et al. previously reported that the 5-year EFS and OS of patients with newly diagnosed ALL in Canada were 89.8% and 94.1% for those enrolled on a therapeutic trial versus 84.1% (p < 0.0001) and 90.5% (p = 0.001) for those not enrolled on a therapeutic trial, respectively [Citation9]. In our cohort of rare and unfavorable B-ALL subtypes, the outcomes of infant ALL were significantly inferior compared to those of hypodiploid and MPAL. In multivariable analyses, WBC ≥50 × 109/L remains the only independent adverse factor for both EFS and OS, irrespective of their underlying ALL subtypes.
The prevalence of infant ALL in Canada is similar to the reported prevalence of 2.5–5% in the literature [Citation4,Citation10]. Despite chemotherapy intensification and incorporation of novel agents, the outcomes of infant ALL remain dismal. The 6-year EFS in the most recently completed infant ALL trial (2006–2016), Interfant-06, was 46.1 ± 2.4% [Citation11], which is higher than the 5-year EFS of 38.9 ± 4.6% in our infant ALL cohort in Canada (2001–2018), but still unacceptably low in comparison to other ALL subtypes. We hypothesize that this survival difference could be partly explained by the retrospective nature of the study which included patients diagnosed in the early era of 2001 where MRD-stratified protocols and enhanced supportive care measures have not been widely adopted. Nevertheless, the era of treatment (prior versus after 2010) was not found to be significant in the univariate analysis of EFS and in the multivariate analysis of OS. Whereas KMT2A status and MRD were identified as the strongest predictors of outcome in infant ALL, our analysis was limited by the paucity of KMT2A status information (only available for 41% of the cohort) and the complete lack of MRD data. In the Interfant-06 study, high-risk patients were defined as infants with a KMT2A rearrangement, less than 6 months of age at diagnosis, and presenting a WBC ≥300 × 109/L or poor prednisone response; the 6-year EFS for these patients was 20.9% [Citation11]. In our infant ALL cohort, multivariable analyses for both EFS and OS identified age <3 months and WBC ≥100 × 109/L at diagnosis were independent adverse prognostic factors. The role of HSCT in the first remission in infant ALL has been explored in predecessor trials [Citation11,Citation12], but the benefits remain controversial. In our cohort, 16 patients (11.5%) underwent HSCT before relapse and this was not found to be a significant predictor in the univariate analyses of EFS and OS.
The prevalence of hypodiploid ALL in our Canadian cohort is similar to that reported in the literature, in the range of 1–2%. The 5-year EFS and OS of hypodiploid patients in our cohort were 63.0% and 75.2%, respectively, which compared favorably to the reported 5-year EFS of 48% in the Children’s Oncology Group (COG) AALL0031 trial [Citation13]. The largest study of hypodiploid ALL to date reported an estimated 5-year EFS and OS of 55.1% and 61.2%, respectively [Citation5]. While early MRD response and modal chromosome distribution (near haploid vs. low hypodiploid vs. high hypodiploid) represent important prognostic factors in previous studies, our analysis did not identify any significant outcome predictor. This could be explained by the identification of hypodiploid ALL patients via ICD-O codes without access to complete cytogenetic data for validation and to further refine classification based on the exact chromosomal number or DNA index, in addition to the absence of MRD information. Moreover, it is possible that our study lacks the statistical power to identify multiple predictors of relapse or death given the small patient number. The COG and other groups reported that HSCT in the first remission did not improve survival in hypodiploid ALL [Citation5,Citation14]. In our cohort, 11 (25.6%) patients with hypodiploid ALL underwent HSCT prior to relapse, and this was not found to be a significant predictor of outcome.
The prevalence of MPAL in our cohort lies toward the lower limit of the reported prevalence range of 1–5% in the literature. The variation in MPAL’s prevalence might be explained by the fact that we did not include T-ALL or AML patients who could fit the criteria for MPAL and conversely, some MPAL cases may not meet the current WHO definition. Furthermore, the lower prevalence in our cohort could also be caused by the difficulty in accurately diagnosing this ALL subtype. The 5-year EFS of MPAL patients in our cohort was 53 ± 9.0%, which was inferior to that of a recent MPAL report with a 5-year EFS of 72 ± 8.0% [Citation6]. It has been demonstrated that ALL-type primary therapy for children with MPAL led to superior survival compared to AML-based or combined-type treatment [Citation15]. Indeed, overall treatment with AML-based protocols was independently associated with worse EFS and OS in our MPAL cohort. In addition, there was a trend toward worse EFS with older age ≥10 years, which is in keeping with the results of a systematic review confirming that adults with MPAL fared worse than their pediatric counterparts [Citation16].
Our study has several limitations that need to be acknowledged. The first of them is its retrospective nature, which is more prone to several biases and a lack of important clinical variables. For instance, our analysis is limited by the available data collected in the registry which did not contain comprehensive cytogenetics information and MRD, both of them being key outcome predictors in the modern era. Consequently, the missing information on the exact modal number of chromosomes or the association of concomitant cancer predisposition syndromes (i.e. Li-Fraumeni) precluded more in-depth analysis among hypodiploid ALL patients. Another limitation was the identification of patients by ICD-O codes. Since the database we used included patients from before 2010, some ICD-O codes were too recent to identify them, notably the codes for hypodiploid ALL and MPAL, the latter has been reclassified in the revised 2016 World Health Organization (WHO) classification of lymphoid neoplasms. To assess the accuracy of MPAL classification in the CYP-C database, we performed an internal review of MPAL cases identified by ICD-O codes at 2 sites and 10/12 cases met the WHO definition for MPAL. The positive predictive value to diagnose MPAL using the ICD-O code was 83.3%, which is acceptable for a national cancer registry. Moreover, we extrapolated HSCT in first remission by interrogating the date of HSCT and the date of relapse, which is not as reliable as it would have been if HSCT in first remission had been specifically collected in the original data collection form. Lastly, AML-based overall treatment was found to be an independent adverse outcome predictor for MPAL in our analysis. However, this may introduce a selection bias since patients who were initially treated with ALL therapy may switch to AML-directed therapy due to poor response or refractory disease. The reasons for switching therapy were not always available for complete and accurate disease status assessment.
In conclusion, infant ALL, hypodiploid ALL, and MPAL represent 8.6% of childhood B-ALL in Canada and are associated with poor prognosis. In the absence of comprehensive molecular and MRD information, WBC ≥50 × 109/L at diagnosis conferred an independent adverse impact on the outcome, irrespective of their underlying ALL subtypes. Our study contributes to establishing the baseline characteristics and outcomes of three rare B-ALL subtypes in Canada and serves as historical controls to design future prospective clinical trials. While the long-term survival of childhood ALL in Canada exceeds 90% with modern chemotherapy regimens, outcomes of our entire cohort remain dismal and highlight the urgent need to harness international collaborations to develop novel treatment strategies for these three rare and high-risk ALL subtypes.
Ethical approval
The study was approved by the Research Ethics Board of CHU Sainte-Justine. The requirement for informed consent and assent was waived given the retrospective nature of this study.
Acknowledgments
The authors gratefully acknowledge the contributions of study participants, participating pediatric oncology centers, members of the Cancer in Young People in Canada (CYP-C) Management and Steering Committees, the Paediatric Oncology Group of Ontario (POGO), and the five POGO Hospital Partners. The CYP-C is fully funded by the Public Health Agency of Canada. We wish to thank all the data managers at the 17 CYP-C sites for their dedicated work in maintaining the CYP-C data quality.
Disclosure statement
THT received consulting fees from Servier and Jazz Pharmaceuticals. The other authors have no conflict of interest to disclose. The analyses and interpretations presented in this work do not necessarily reflect the opinions of the government of Canada.
Data availability statement
Data used in this publication are from the Cancer in Young People in Canada Surveillance Program and are used with the permission of the Public Health Agency of Canada.
References
- Hunger SP, Mullighan CG. Acute lymphoblastic leukemia in children. N Engl J Med. 2015;373(16):1541–1552.
- Hunger SP, Loh ML, Whitlock JA, et al. Children’s oncology group’s 2013 blueprint for research: acute lymphoblastic leukemia. Pediatr Blood Cancer. 2013;60(6):957–963.
- Raetz EA, Bhatla T. Where do we stand in the treatment of relapsed acute lymphoblastic leukemia? Hematology Am Soc Hematol Educ Program. 2012;2012:129–136.
- Biondi A, Cimino G, Pieters R, et al. Biological and therapeutic aspects of infant leukemia. Blood. 2000;96(1):24–33.
- Pui CH, Rebora P, Schrappe M, et al. Outcome of children with hypodiploid acute lymphoblastic leukemia: a retrospective multinational study. J Clin Oncol. 2019;37(10):770–779.
- Orgel E, Alexander TB, Wood BL, et al. Mixed-phenotype acute leukemia: a cohort and consensus research strategy from the children’s oncology group acute leukemia of ambiguous lineage task force. Cancer. 2020;126(3):593–601.
- Public Health Agency of Canada. Cancer in Young People in Canada: a report from the enhanced childhood cancer surveillance. Winnipeg (Canada): Public Health Agency of Canada; 2017. p. 58. https://www.canada.ca/fr/sante-canada/services/publications/science-recherche-et-donnees/cancer-jeunes-canada-surveillance-2017.html
- Mitra D, Hutchings K, Shaw A, et al. Status report - the cancer in young people in Canada surveillance system. Health Promot Chronic Dis Prev Can. 2015;35(4):73–76.
- Strahlendorf C, Pole JD, Barber R, et al. Enrolling children with acute lymphoblastic leukaemia on a clinical trial improves event-free survival: a population-based study. Br J Cancer. 2018;118(5):744–749.
- Brown P. Treatment of infant leukemias: challenge and promise. Hematology Am Soc Hematol Educ Program. 2013;2013:596–600.
- Pieters R, De Lorenzo P, Ancliffe P, et al. Outcome of infants younger than 1 year with acute lymphoblastic leukemia treated with the interfant-06 protocol: results from an international phase III randomized study. J Clin Oncol. 2019;37(25):2246–2256.
- Pieters R, Schrappe M, De Lorenzo P, et al. A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (interfant-99): an observational study and a multicentre randomised trial. Lancet. 2007;370(9583):240–250.
- Schultz KR, Devidas M, Bowman WP, et al. Philadelphia chromosome-negative very high-risk acute lymphoblastic leukemia in children and adolescents: results from children’s oncology group study AALL0031. Leukemia. 2014;28(4):964–967.
- McNeer JL, Devidas M, Dai Y, et al. Hematopoietic stem-cell transplantation does not improve the poor outcome of children with hypodiploid acute lymphoblastic leukemia: a report from children’s oncology group. J Clin Oncol. 2019;37(10):780–789.
- Hrusak O, de Haas V, Stancikova J, et al. International cooperative study identifies treatment strategy in childhood ambiguous lineage leukemia. Blood. 2018;132(3):264–276.
- Maruffi M, Sposto R, Oberley MJ, et al. Therapy for children and adults with mixed phenotype acute leukemia: a systematic review and meta-analysis. Leukemia. 2018;32(7):1515–1528.