PDF
Citation
Print
Introduction
Over the past several decades, the survival rate of pediatric patients with acute lymphoblastic leukemia (ALL) has significantly improved. A recent study reports the overall survival (OS) rate for pediatric ALL exceeding 90% [1]. Conventional chemotherapeutic agents were developed before 1970s, leading to the introduction of combination chemotherapy schedules based on evaluations of tolerability and treatment response [2]. Since then, customized treatment has gradually evolved, incorporating risk stratification based on each patient’s clinical factors, characteristics of the leukemic blasts, and response to early therapy, leading to improved survival rate of pediatric ALL [3]. Multicenter clinical trials have been instrumental in optimizing treatment effectiveness and reducing side effects. This progress has allowed for a reduction in radiation therapy and the determination of optimal chemotherapy doses and treatment durations [4-9]. Additionally, individual genetic factors and pharmacokinetics now play crucial roles in treatment decisions, and molecular-targeted agents and immunotherapy have emerged as novel therapeutic strategies over the last decade [2].
At our institution, the treatment regimen for ALL has evolved systematically over the last two decades, guided by considerations of risk stratification, and treatment response. In this study, we report our 20-year experience with 460 children and adolescents with ALL who were treated with a regimen tailored to their risk stratification.
Materials and Methods
1. Patients
This study included patients diagnosed with ALL and treated at the Pediatrics Department of Samsung Medical Center (SMC) between January 2000 and December 2019.
2. Risk stratification
The National Cancer Institute (NCI)–Rome criteria define standard-risk (SR) as age < 10 years and a white blood cell (WBC) count of < 50×109/L, and high-risk (HR) as age ≥ 10 years and/or a WBC count ≥ 50×109/L.
At SMC, patients aged 10 years or older at diagnosis or with an initial WBC count ≥ 50×109/L were classified as HR patients. Additionally, patients with a T-cell immunophenotype, mixed phenotype, or a history of previous steroid treatment were also categorized as HR. Patients exhibiting at least one of the following characteristics—an initial WBC count ≥ 200×109/L, positive BCR::ABL1 rearrangement, hypodiploidy (< 45 chromosome), t(4;11), or failure of remission induction—were considered very high-risk (VHR). All other patients were classified as SR patients.
3. Treatment scheme
The induction therapy followed the risk group definitions from the NCI-Rome criteria.
In the SR group, the treatment protocol was based on a modified Children’s Cancer Group (CCG)-1891 regimen [9]. The remission-induction protocol included prednisolone, L-asparaginase, and vincristine with intrathecal chemotherapy. Doxorubicin was added to the HR group.
The SMC protocol for the HR group underwent two modifications during the study period. From 2000 to 2005, a modified CCG-1882 protocol was used; from 2006 to 2014, the Korean multicenter HR ALL-0601 protocol was implemented, and from 2015 to 2019, the Korean multicenter HR ALL-1501 protocol was adopted (Fig. 1). In a modified CCG-1882 protocol and the Korean multicenter HR ALL-0601 protocol, the inclusion criteria was based on SMC risk stratification. However, in the Korean multicenter HR ALL-1501 protocol, patients with an initial WBC count ≥ 200×109/L, and patients with central nervous system (CNS) leukemia and testicular leukemia were included. In contrast, early T-cell precursor leukemia was excluded. The modified CCG-1882 protocol featured an augmented Berlin-Frankfurt (BFM) regimen with extended and intensified post-induction phases [10]. Initial consolidation comprised a six-agent chemotherapy combination with 18 Gy cranial irradiation (CRT) for patients without CNS disease at diagnosis, while patients with CNS disease at diagnosis received 24 Gy of cranial and 6 Gy of spinal irradiation. The Korean multicenter study-0601 protocol, based on the CCG protocol backbone (Table 1), included bone marrow examination on the 7th and 14th days of induction. Based on the results, patients were categorized as rapid early responders (RER) with less than 5% blasts (M1) on the 7th-day test or 5%-25% blasts (M2) on the 7th-day and M1 on the 14th-day test. In contrast, patients with more than 25% blasts (M3) on the 7th-day test or with M2 or M3 on the 14th-day test were categorized as slow early responders (SER). In cases of CNS3 or testicular leukemia, patients were classified under the SER category. The RER received two courses of interim maintenance (IM) and one course of delayed intensification (DI), while the SER received double IM and DI, along with prophylactic CRT (12 Gy) during the second DI phase. Patients with CNS disease at diagnosis received two additional intrathecal methotrexate (MTX) injections during induction, followed by CRT (18 Gy) and spinal radiotherapy (6 Gy) during the second DI. Treatment was continued for 2 years for girls and 3 years for boys, from the first IM to the final maintenance therapy. In the Korean multicenter study-1501 protocol, aiming at reducing long-term complications, prophylactic radiation therapy was replaced with intrathecal treatment with MTX, cytarabine, and hydrocortisone three-agent therapy. Two cycles of IM therapy with high-dose intravenous MTX were also introduced to effectively prevent CNS relapse.
Maintenance therapy with vincristine/steroid pulses was administered every 12 weeks, with the duration of treatment set at 3 years for males and 2 years for females from the first IM therapy.
The induction protocol for the VHR group was identical to that used for the HR group. Patients with Philadelphia chromosome were continuously given Imatinib at a dose of 340 mg/m2/day from diagnosis. The consolidation protocol for the VHR group is outlined in S1 Table. After the third cycle of consolidation, allogeneic hematopoietic stem cell transplantation was performed. Infant patients under 1 year of age were treated with a modified CCG protocol (Table 1), and if MLL rearrangement was present, they underwent hematopoietic stem cell transplantation after consolidation therapy (S1 Table). In patients with Down syndrome ALL, the difference in the treatment protocol was that all induction therapy was administered with three drugs. Only if the bone marrow on day 14 was M2 or M3, daunorubicin at 25 mg/m² was added on days 21 and 28. Following consolidation therapy, one cycle of intermediate maintenance therapy with intravenous MTX at 1,000 mg/m2 was given, along with one cycle of reintensification therapy. Intrathecal injections during induction therapy were done with intrathecal MTX, while subsequent intrathecal therapy used a triple-drug regimen. Leucovorin rescue was administered after every intrathecal MTX injection. Additionally, the duration of maintenance therapy was set at 2 years for both males and females, starting from the first IM therapy. In cases of mixed phenotype acute lymphoblastic leukemia, the HR treatment protocol for ALL was followed. However, if complete remission (CR) was not achieved, the treatment protocol was switched to the acute myeloid leukemia protocol.
Patients who experienced a relapse underwent reinduction, after which they either continued chemotherapy or underwent allogeneic hematopoietic stem cell transplantation, depending on their risk stratification.
4. Statistical analysis
Kaplan-Meier survival analysis was conducted to estimate the probabilities of relapse-free survival (RFS) and OS rates. OS was defined as the time from the diagnosis of ALL to death from any cause or the last visit, while RFS was defined as the time from diagnosis to an event recognized as relapse (defined as the infiltration of marrow by more than 5% blast cells in previously normal bone marrow and/or an emergence of extramedullary leukemia) or death due to any cause, whichever occurred first. Differences in OS and RFS between groups and clinical factors were compared using the log-rank test, with a p-value < 0.05 considered statistically significant. All statistical analyses were performed using SPSS ver. 18 (SPSS Inc.).
Results
1. Patient characteristics
Between January 2000 and December 2019, 549 patients were diagnosed with or treated for ALL at SMC. Out of these, 89 were excluded from the study for various reasons: 74 were transferred from another hospital after diagnosis, 13 were transferred to another hospital before treatment, and two passed away before chemotherapy.
Among the remaining 460 patients, 250 were male and 210 were female, with a median age at diagnosis of 5.5 years (range, 0.1 to 26.5 years). The median initial WBC count was 9.330×109/L (range, 0.31 to 767.44). According to the NCI definition for risk group classification, 248 patients were in the SR group, and 212 patients were in the HR group. Regarding immunophenotype, 385 patients exhibited the B-cell type, 39 showed the T-cell type, and 36 had a mixed phenotype. Cytogenetic testing revealed 39 patients with VHR features including Philadelphia chromosome in 24 patients, t(4;11) in 12 patients, and hypodiploidy in three patients. Additionally, 123 patients had normal chromosomes, 100 had hyperdiploidy (chromosomal number 51-65), and 68 had ETV6::RUNX1 abnormality. In the SMC risk stratification, with consideration on the above findings, 213 patients were classified as SR, 131 as HR, and 51 as VHR. The clinical characteristics of the patients are summarized in Table 2.
2. Treatment
The median follow-up period was 97 months (range, 1 to 253 months). Among the 460 patients, 191 were treated with a modified CCG-1891 protocol, 35 with a modified CCG-1882 protocol, 96 with protocol 0601, and 53 with protocol 1501, while 85 were treated with other protocols.
At the end of induction, 17 patients failed to achieve CR and seven died due to complications during the induction treatment. Hematopoietic stem cell transplantation was performed in 126 patients, with transplantation rates according to the changed protocol being 54.3% (19 out of 35 patients), 36.5% (35 out of 96 patients), and 30.2% (16 out of 53 patients) in the modified CCG-1882, 0601, and 1501 protocol treatment groups, respectively (p=0.067). Among the HR patients treated with the 0601 and 1501 protocols, 46 patients were classified into the RER group, and 41 patients were classified into the SER group based on bone marrow evaluation on induction day 7 and day 14. The remaining patients did not meet the evaluation criteria.
3. Survival and prognostic factors
The 10-year OS was 83.8%±1.9% and the 10-year RFS was 77.3%±2.1% for all patients (Fig. 2A and B). Survival rates based on clinical factors are presented in Table 3. Statistically significant differences were observed concerning age at diagnosis, initial WBC counts, and immunophenotype. However, no significant differences were found in sex and extramedullary involvement. Regarding age at diagnosis, the 10-year survival rate was lowest for those under 1 year of age (RFS, 60.0%±11.0%; OS, 64.6%±10.8%), while it was highest for those between 1 and 10 years old (RFS, 80.5%±2.4%; OS, 89.4%±1.8%). For individuals aged 10 years or older, the 10-year RFS was 72.4%±4.3% and the OS was 72.6%±4.5%. Notably, the prognosis was particularly poor for those aged 15 years and older, with only a 56.5%±13.1% survival rate (Fig. 3A). According to the NCI risk group, the 10-year RFS was 82.3±2.7% in the SR group and 71.5%±3.3% in the HR group (p < 0.001). The 10-year OS was 92.8%±1.8% in the SR group and 73.1%±3.3% in the HR group (p < 0.001) (Fig. 3B).
The survival rate based on cytogenetic groups also demonstrated a statistically significant difference. The 10-year OS for patients with unfavorable cytogenetics was 57.0%±7.9%, with t(4;11) at 66.7%±13.6%, and the Philadelphia chromosome at 58.0%±10.1%. The 10-year OS of patients with hypodiploidy was the lowest at 33.3%±27.2%. In contrast, the OS was 81.2%±3.6% for cases with normal chromosomal findings, 90.7%±3.7% for ETV6::RUNX1 abnormality, and reached 93.1%±3.6% for cases with hyperdiploidy, indicating the most favorable outcomes (Fig. 3D). Out of 23 infant ALL patients, 17 had MLL rearrangements including t (4;11). The 10-year RFS and OS rates for these patients were 52.3%±14.6% and 57.8%±12.2%, respectively. In comparison, the RFS and OS rates for patients without MLL rearrangements were both 66.7%±19.2%. Although the survival rates were lower for patients with MLL rearrangements, the difference was not statistically significant (p=0.597 and p=0.794). According to the SMC risk groups, reflecting both cytogenetics and immunophenotype, the 10-year RFS was 84.7%±2.8% in the SR group, 83.0%±3.5% in the HR group, and 63.5%±6.8% in the VHR group. The OS for these groups were 95.9%±1.4%, 83.8%±3.6%, and 66.2%±6.9%, respectively (Fig. 3C).
The survival rate for each protocol was considered as the 5-year survival rate due to the shorter follow-up duration of the 1501 protocol. In the SR group the 5-year RFS for the modified CCG-1891 protocol was 90.5%±2.2%, and the OS was 96.8%±1.3%. For the HR group, the RFS for the different protocols were as follows: modified CCG-1882 protocol, 65.5%±8.1%; 0601 protocol, 77.7%±4.3%; 1501 protocol, 90.6%±4.8% (p < 0.001). The OS for these groups were 73.9%±7.5%, 83.0%±3.9%, and 96.2%±2.6%, respectively (p < 0.001) (Fig. 4A). Among the HR patients treated with the 0601 and 1501 protocols, those classified as RER and SER had the following outcomes: The 5-year RFS was 90.2%±4.7% in the RER group and 85.5%±6.2% in the SER group (p=0.109), and OS was 93.1%±3.8% in the RER group and 94.9%±3.6% in the SER group (p=0.264). By protocol, in the 0601 protocol, RFS were 90.0%±5.5% and 83.3%±7.6% for the RER and SER groups, respectively (p=0.113), while in the 1501 protocol, RFS were 92.9%±6.9% and 88.9%±10.5%, respectively (p=0.875). The OS were 93.3%±4.6% for the RER group and 91.7%±5.6% for the SER group in the 0601 protocol (p=0.062). In the 1501 protocol, OS for the RER was 93.3%±6.9%, while there were no reported mortalities in the SER group (p=0.270). These results indicate comparable survival rates with no inferiority observed in the SER group (Fig. 4B).
4. Relapse and treatment related mortality
The survival rate of relapsed patients was poor. The 5-year OS of 71 relapsed patients was 37.7%±6.0% (Fig. 5A). The 10-year cumulative incidence of relapse for all patients was 17.3%±2.0% (Fig. 2C). Among the 71 relapsed patients, 51 experienced bone marrow relapse, 12 had extramedullary relapse, and eight had both. In terms of timing, 59 patients experienced relapse within 60 months of diagnosis, while 12 patients had very late relapse occurring more than 60 months after diagnosis (Table 4). This very late relapse group included nine males and three females, one of whom had Down syndrome. The average time to relapse was 89 months (range, 60 to 143 months) from diagnosis. There were seven cases of bone marrow relapse, two cases of isolated extramedullary relapse (one in the CNS and one in the testis), and three cases of both. The extramedullary relapse rate was 41.6% in the very late relapse group, higher than the group that relapsed before 60 months (17 out of 59 patients, 28.3%), although the difference was not statistically significant (p=0.125). At relapse, one patient had a new genetic aberration with t(8;9;22) and another patient had a new rearrangement of KMT2A. Of the 12 patients, seven received chemotherapy alone and five underwent hematopoietic stem cell transplantation, whereas among earlier relapse patients, 47 out of 59 underwent hematopoietic stem cell transplantation (p=0.007). The 5-year OS was 71.3%±14.1% in the very late relapse group (> 60 months) and 31.5%±6.3% in those relapsed earlier (p=0.050) (Fig. 5B).
The 10-year treatment related mortality (TRM) rate in all patients was 9.0%±1.5% (Fig. 2D). There were a total of 36 cases of TRM, with the causes as follows: 17 due to infection, eight due to organ failure, four due to bleeding, six due to graft-versus-host disease, and one due to tumor lysis syndrome. Additionally, there were six cases of secondary malignancy, including myelodysplastic syndrome, post-transplant lymphoproliferative disorder (PTLD), skin lymphoma, papillary thyroid carcinoma, rhabdomyosarcoma, and optic tract spindle cell tumor. Among these cases, only the PTLD case resulted in mortality due to infection.
Discussion
Years of clinical studies have consistently demonstrated that optimal treatment outcomes for pediatric ALL are achieved through risk stratification and treatment adjustments based on prognostic factors [11]. Most centers stratify patients with pediatric ALL into the SR, HR, and VHR groups, applying distinct treatments. A recent Children’s Oncology Group (COG) clinical study reported a 5-year survival rate of 90%-95% or higher for the SR group and 75%-90% for the HR group [12,13]. Various efforts, including protocol changes, have aimed to improve survival rates in the HR group. Over the past two decades at SMC, two treatment protocol changes for HR patients have been implemented.
In our study, patient survival improved over time with protocol changes. The 5-year RFS with the modified CCG-1891 protocol for the SR group was 90.5%±2.2%, with an OS of 96.8%±1.3%. For the HR group, the RFS rates for the different protocols were as follows: modified CCG-1882 protocol, 65.5%±8.1%; 0601 protocol, 77.7%±4.3%; 1501 protocol, 90.6%±4.8%. The OS rates for these groups were 73.9%±7.5%, 83.0%±3.9%, and 96.2%±2.6%, respectively. This aligns with recent ALL studies [4-7,14,15]. Advances in diagnostic technology and enhanced supportive care have also contributed, but further optimization of treatment protocol have had the most significant impact on the improved survival rates [16].
Changes to the HR group protocol included modifications to the inclusion criteria based on risk stratification using prognostic factors. Specifically, under the previous protocol, hyperleukocytosis with a WBC count ≥ 200×109/L was included in the VHR group. However, starting with the 1501 protocol, this upper limit of WBC count was included in the HR group inclusion criteria. VHR is defined as cases with a projected 5-year event-free survival (EFS) rate of less than 45% when treated with chemotherapy alone. The Children’s Cancer Group reported a 5-year EFS of 58% for cases with a WBC count ≥ 200×109/L, indicating that this factor alone does not categorize as VHR [8]. Similarly, patients with CNS leukemia or testicular involvement at diagnosis were categorized into the HR group, as these conditions are associated with poor treatment outcomes under SR therapy [8]. Additionally, recent studies have shown that early precursor T-cell acute lymphoblastic leukemia exhibits poor survival rates with conventional chemotherapy alone, necessitating a more intensive therapeutic approach [17]. Therefore, this type of leukemia was excluded from the HR group in the 1501 protocol.
Furthermore, the protocol changes focus on individualizing treatment intensity and duration based on the initial treatment response, avoiding prophylactic cranial radiation and determining accurate prognostic factors. These important strategies were taken into account in changing the treatment protocols for our patients. Rapid response to remission induction chemotherapy has proven to be a crucial prognostic factor in many studies [18,19]. Therefore, efforts have been made to improve treatment outcomes of SER through more intensified treatments. In a prospective CCG study involving 311 HR pediatric ALL patients showing slow early response, those who received intensified post-remission chemotherapy demonstrated a superior survival rate than those who did not (5-year EFS 75.0%±3.8% vs. 55.0%±4.5%, p < 0.001) [10]. In line with these findings, another Korean single-center study demonstrated that, in HR pediatric ALL patients, intensified post-remission treatments for the SER group within the structure of the existing CCG-1882 protocol effectively reduced the survival gap between RER and SER patients (5-year EFS rate, 85.4% vs. 72.7%; p=0.212) [20]. These findings suggest that the poor outcome of slow early response could be overcome by treatment augmentation following remission induction. As observed in our cohort, recent protocol changes led to improved survival outcomes, largely affected by post-induction intensification for patients with SER.
Recent studies have explored the appropriate use and dosage of MTX in SER. CNS disease, a major cause of treatment failure, was included in the HR group in our study, highlighting its significance. MTX plays an important role in preventing CNS relapse. Two different MTX intensification strategies have been studied. One is the Capizzi regimen, in which MTX doses are escalated sequentially from 100 to 300 mg/m2 without leucovorin rescue, followed by asparaginase injection. Another one is the high dose–MTX (HD-MTX) regimen, in which 2-5 g/m2 of MTX is administered over 24 hours, followed by leucovorin rescue [10,21]. The Capizzi regimen used to be the key element of the augmented BFM protocol in previous HR group studies [22]. However, the COG AALL0232 study favored HD-MTX over the Capizzi regimen in children and young adults with HR ALL (5-year EFS, 82% vs. 75.4%; p=0.006), showing superior efficacy without increased toxicity. Notably, HD-MTX decreased both marrow and CNS recurrences [5,23]. This COG study also demonstrates that intensified post-remission treatments can successfully replace prophylactic CNS irradiation. Similarly, our 1501 protocol aimed to reduce long-term complications by omitting prophylactic CNS irradiation, enhancing intrathecal treatment, and introducing HD-MTX for SER. Despite a relatively short follow-up, promising results were seen with a 5-year RFS of 90.6%±4.8% and OS of 96.2%±2.6%.
Age significantly affected survival rates (p < 0.001), with the best prognosis for patients aged 1-10 years. Patients under 1 year and over 10 years had poorer outcomes. In particular, the survival rate of patients aged over 15 years was as low as those under 1 year. The unique biology of ALL in adolescents and young adults (AYA) poses challenges. Recurrent cytogenetic abnormalities associated with a good prognosis, including hyperdiploidy and ETV6::RUNX1, are less frequently found in patients with AYA than in younger children. In contrast, BCR::ABL1, associated with poor prognosis, is more frequently observed with increasing age. Additionally, amplification within chromosome 21 (iAMP21) and IKZF1 gene alteration, both known to be associated with poor outcomes, are more commonly found in adolescents than younger children [24,25]. Both the United Kingdom Acute Lymphoblastic Leukaemia (UKALL) 2003 and the Nordic Society of Pediatric Hematology and Oncology (NOPHO) ALL 2008 trials showed higher levels of post-induction minimal residual disease (MRD) in AYAs compared with younger children [14,26]. Further studies are needed to clarity the detailed characteristics in this age group.
In this study, 71 patients (15.4%) experienced recurrence, consistent with major pediatric clinical studies. The 5-year OS of relapsed patients was only 37.7%±6.0%. Interestingly, the OS of the St. Jude Total Therapy Study XVI was similar to that of the previous Total Therapy Study XV, suggesting that further treatment intensification is not likely to improve the outcomes [2,7]. Novel therapeutic approaches including targeted agents may be necessary to further improve the outcomes of pediatric patients with ALL.
Among the 71 relapsed patients, 12 experienced late relapses occurring 5 years after diagnosis, with the longest being 143 months. A short first remission duration has been associated with a significantly worse prognosis. On the other hand, late relapses may need only moderately intensified treatments [27,28]. However, some studies revealed that late relapsers with new cytogenetic findings distinct from their initial diagnosis tend to have poor outcomes, requiring more intensive treatments [29,30]. In our cohort, some late relapsers exhibited new mutations, such as t(8;9;22) and KTM2A. However, despite a higher percentage of late relapse patients being treated with chemotherapy alone rather than undergoing hematopoietic stem cell transplantation, compared to those who relapsed within 60 months, there was no statistically significant difference in the 5-year OS between the two groups (5-year OS rate, 71.3%±14.1% vs. 31.5%±6.3%; p=0.05). These findings underscore the need for continued research to identify specific prognostic markers and personalized therapeutic strategies for late relapsers based on genetic profiling. Given the rarity of very late relapsers, long-term follow-up and multicenter collaboration is crucial to identify prognostic factors for this group.
In conclusion, our findings underscore the positive impact of treatment amendments guided by risk stratification on clinical outcomes. Despite the significant progress made, challenges persist for patients with unfavorable prognostic factors and relapsed disease. Addressing these challenges calls for multicenter collaborative approaches to collectively overcome current hurdles. A major limitation of our study is the lack of reliable MRD test results, which are crucial for risk stratification and could significantly influence treatment outcomes. Nevertheless, the recent availability of next-generation sequencing MRD tests in Korea holds promise for providing valuable insights into risk stratification and treatment adaptation in future multicenter collaborative clinical trials.
XML Download