|Home | About | Journals | Submit | Contact Us | Français|
The optimal form of treatment for children with relapsed or refractory acute promyelocytic leukemia (APL) is unclear. We retrospectively analyzed the results of 32 (11 autologous, 21 allogeneic) hematopoietic stem cell transplants (HSCT) performed for children originally treated on either the Eastern Cooperative Group E2491 Trial or the Cancer and Leukemia Group B C9710 Trial and subsequently diagnosed with relapsed or refractory APL. For autologous HSCT, the incidence of treatment-related mortality (TRM) and relapse were 0% (95% CI, 0-30%) and 27% (95% CI, 9-57%), respectively. The 5-year event-free survival (EFS) and overall survival (OS) following autologous HSCT was 73% (95% CI, 43-91%) and 82% (95% CI, 51-96%), respectively. For allogeneic HSCT, the incidence of TRM and relapse were 19% (95% CI, 7-41%) and 10% (95% CI, 2-30%) respectively. The 5-year EFS and OS following allogeneic HSCT was 71% (95% CI, 50-86%) and 76% (95% CI, 55-90%) respectively. There was no significant difference in EFS or OS between autologous and allogeneic HSCT. This data demonstrates that autologous and allogeneic HSCT are both effective therapies for treatment of children with relapsed or refractory APL. Autologous HSCT is associated with a low incidence of TRM, while allogeneic HSCT is associated with a low incidence of relapse, suggesting a strong GVL effect against residual APL.
In children, acute promyelocytic leukemia (APL) is a rare subtype of acute myeloblastic leukemia (AML), accounting for only 4-8% of AML in the USA (1). All cases are associated with an abnormal fusion of the retinoic acid receptor (RAR)-α gene on chromosome 17, usually to the promyelocytic (PML) gene on chromosome 15 (1). Since the introduction of all-trans retinoic acid (ATRA) combined with conventional therapy, the disease-free survival of children with newly diagnosed APL now approaches 75% (2-6).
For those patients with relapsed APL, occasional patients have demonstrated second remissions for as long as 8 years following treatment with chemotherapy and/or ATRA (7-10). However, most patients ultimately undergo consolidative autologous(10-14), allogeneic (10-13, 15, 16), or haploidentical hematopoietic stem cell transplantation (HSCT) (17). The majority of reports on the use of HSCT for treatment of relapsed or refractory APL deal primarily with adult patients, making the benefit of these therapies for children unclear. Here, we report the outcomes in 32 children originally treated on the Eastern Cooperative Group (ECOG) E2491 Trial (6, 18, 19) or the Cancer and Leukemia Group B (CALGB) C9710 Trial(20) from 1992-2005 who underwent any form of HSCT for treatment of relapsed or refractory APL.
Approval for the study was obtained by the ECOG and CALGB Executive Review Committees. The 32 patients included on this study were diagnosed with APL according to bone marrow morphology (19). Cytogenetic analysis and/or reverse-transcriptase polymerase chain reaction (RT-PCR) confirmed the presence of t(15;17) in all cases.
Between May 1992 and February 1995, 71 of the 350 patients enrolled on the ECOG E2491 trial were children (ages 1-18). Patients were randomly assigned to receive induction therapy with either daunorubicin (45 mg/m2/day for 3 days) plus cytarabine (100 mg/m2/day for 7 days), or ATRA (45 mg/m2/day until complete remission or 90 days). Two cycles of consolidation chemotherapy with daunorubicin and cytarabine were administered, following which patients were randomly assigned to receive maintenance ATRA for one year or observation. The results of the ECOG E2491 trial have been previously reported (6, 18, 19).
Between June 2000 and March 2005, 91 of the 582 patients enrolled on the CALGB C9710 trial were children (ages 1-18). Patients were treated with ATRA (45 mg/m2/day from day 1 until achievement of CR), cytarabine (200 mg/m2/day for 7 days from day 3 to day 9), and daunorubicin (50 mg/m2 for 4 days; 1.5 mg/kg for 4 days for age <3). Patients were randomly assigned to either proceed directly to two cycles of consolidation with ATRA plus daunorubicin, or to first receive two cycles of arsenic trioxide (ATO). Patients were then randomized to receive maintenance therapy with ATRA alone for 1 year, or ATRA plus 6-mercaptopurine and methotrexate for 1 year. The results of the CALBG C9710 trial have been preliminarily reported (20).
A total of 52 children (ages 1-18 years at the time of diagnosis) treated on these two trials were reported to have relapsed or had primary-refractory disease, defined as failure to achieve complete remission with the protocol induction chemotherapy. Six patients are known to have not undergone HSCT (death during reinduction, n = 2; refused HSCT and died, n = 2; treated with differentiating chemotherapy only, n = 2). Fourteen patients were excluded from the analysis due to an inability to obtain sufficient data on HSCT type and outcomes from the transplant centers. Characteristics at the time of diagnosis of the remaining 32 patients who underwent HSCT are summarized in Table 1.
Multiple treatment regimens were utilized, as shown in Table 2. Nineteen patients received differentiating therapy alone (ATRA, n = 8; arsenic trioxide (ATO), n = 9; ATRA + ATO, n = 2). Eight patients received differentiating agents plus chemotherapy, including daunorubicin, etoposide, idarubicin, and cytarabine, all of which have been shown to have efficacy against APL(13, 21). Three patients were treated with only cytotoxic chemotherapy regimens, the most common of which was idarubicin plus cytarabine. The treatment for 2 patients was unknown. Molecular assessment of the PML-RARα fusion gene by RT-PCR was not performed before HSCT in all patients. When RT-PCR was performed, assay sensitivities varied from 1×10−3 to 1×10−5.
Eleven patients underwent autologous HSCT and 21 patients underwent allogeneic HSCT. RT-PCR of the autograft product was negative in 6 patients and unknown in 5 patients. Multiple different donor types, conditioning regimens, and graft-versus-host disease (GVHD) prophylaxis strategies were utilized, as shown in Table 3. When available, HLA-typing examined either 8 genetic loci (A-, B-, C-, DR-loci) or 10 (A-, B-, C-, DR-, DQ-loci). The stem cell source was known for 16 patients, (bone marrow, n =12; peripheral blood stem cells, n = 3; and cord blood, n =1). Thirteen patients received total body irradiation (TBI) as part of the conditioning regimen, with doses ranging from 1200-1375 cGy. The most commonly employed conditioning regimen utilized a two agent combination of cyclophosphamide with either TBI or busulfan. Most of the other regimens added a third agent (including cytarabine, etoposide, or thiotepa) to that backbone. Three patients who received mismatched transplants had ex vivo T-cell depletion for GVHD prophylaxis.
Categoric variables were compared by two-sided Fisher's exact test. Event-free survival (EFS) and overall survival (OS) were estimated by the Kaplan-Meier method using log-rank tests (SPSS 16.0, Chicago, IL). Events were defined as graft failure, treatment-related mortality (TRM), or relapse. There were no secondary malignancies reported.
As shown in Table 1, 10 patients (32%) were considered high-risk at diagnosis based on a WBC count greater than 10,000 cells/μL(22). Other variables, including the microgranular variant and the presence of additional chromosomal abnormalities, were present in a small number of patients. There were no significant differences in presenting characteristics between patients who underwent autologous vs. allogeneic HSCT.
As shown in Table 2, three patients failed to achieve CR with their initial protocol induction therapy and underwent HSCT for treatment of primary-refractory disease. The other 29 patients underwent HSCT for treatment of relapse, which occurred at a median of 10 months (range, 1-46 months) from time of first remission, with 41% of relapses occurring later than 12 months from time of first remission. All relapses occurred in the bone marrow. Four patients also had evidence of a concomitant extramedullary relapse (cerebrospinal fluid, n =2; skin, n =2). All patients achieved morphologic CR prior to HSCT. Sixteen patients were known to have had RT-PCR analysis for PML-RARα prior to HSCT, 2 of whom were positive (assay sensitivity = 1×10−3 for one patient, 1×10−4 for the other). There were no significant differences in relapse characteristics between patients who underwent autologous vs. allogeneic HSCT.
Patient characteristics at time of HSCT are shown in Table 3. Patients undergoing allogeneic HSCT were more likely to have received a TBI-based conditioning regimen (p = 0.02). Statistical analysis (data not shown) did not reveal a significant impact of age at time of transplant, time from achievement of CR to HSCT, type of allogeneic donor, type of conditioning regimen, or type of GVHD prophylaxis upon any of the endpoints (TRM, relapse, EFS, or OS).
Engraftment data was available on all patients. All but two patients achieved primary neutrophil engraftment (94%; 95% CI, 79-99%). One patient died prior to neutrophil engraftment on day +11 post-allogeneic HSCT secondary to necrotizing tracheobronchitis. One patient who received a bone marrow transplant from a matched related donor (MRD) showed no evidence of WBC recovery on day +28 and was determined to have primary graft failure. A second infusion of peripheral blood stem cells from the same donor was performed with successful engraftment, but the patient died on day +469 secondary to complications of chronic GVHD. Including the 2 patients above, a total of 4 patients (13%; 95% CI, 4-29%) died of transplant-related causes between days +11 and +497 after HSCT. The other causes of TRM were idiopathic pneumonia syndrome, and post-transplantation lymphoproliferative disease. All patients experiencing TRM had received allogeneic transplants (MRD, n =2; partially-matched related donor, n =1; and matched unrelated donor, n =1). Therefore, the 3-year cumulative incidence of TRM following autologous HSCT was 0% (95% CI, 0-30%) and the 3-year cumulative incidence of TRM following allogeneic HSCT was 19% (95% CI, 7-41%). Statistical analysis (data not shown) did not show any factors significantly associated with an impact upon TRM, though all treatment-related deaths took place early in the study period, with none occurring after 1996.
Information on GVHD was available for 16 of the 19 patients who engrafted following allogeneic HSCT, including 8 recipients of related-donor transplants and 8 unrelated-donor transplants. Five of the 16 patients (31%; 95% CI, 14-56%) developed Grade II-IV acute GVHD (MRD graft, n =2; unrelated donor graft, n =3). Two of the 16 patients (13%; 95% CI, 2-37%) developed extensive chronic GVHD.
A total of 5 patients relapsed following HSCT, including 3 patients (27%; 95% CI, 9-57%) who relapsed after autologous HSCT and 2 patients (9%; 95% CI, 2-30%) who relapsed after allogeneic HSCT. Median time to relapse was 15.3 months (range, 4.4-42.7 months) and 13.5 months (range 7.3-19.6 months) following autologous and allogeneic HSCT, respectively. Both patients with positive RT-PCR for the PML-RARα fusion transcripts prior to HSCT were in long-term molecular remission (2nd CR duration 14 or more months longer than 1st CR) post-HSCT (1 autologous and 1 allogeneic) at last follow-up.
Of note, all 3 patients who relapsed after autologous HSCT had a negative RT-PCR for the PML-RARα fusion transcript prior to HSCT and 2 had a RT-PCR negative autograft product (1 was not reported). Of these 3 patients, 1 died from progressive APL 40 days after relapse (1323 days post HSCT), and 2 received second transplants from allogeneic sources. One of the second transplant patients died following an unrelated donor transplant from a disseminated adenoviral infection and the other survived in long-term molecular remission after a matched-sibling cord blood transplant (previously reported) (17).
The 2 relapses that occurred after allogeneic HSCT both occurred in recipients of MRD grafts, in whom no acute or chronic GVHD occurred. Both patients received second transplants. One patient died of invasive aspergillosis following a second transplant from the same MRD, and 1 patient survived in long-term molecular remission after a paternal haploidentical transplant (previously reported) (17).
As shown in Figure 1, for the 32 patients originally treated on the ECOG E2491 or CALGB C9710 trials who underwent HSCT for treatment of relapsed or refractory APL, the 5-year EFS was 73% (95% CI, 43-91%) with autologous HSCT, and 71% (95% CI, 50-86%) with allogeneic HSCT. One patient in each group was successfully rescued with second HSCT following relapse after first HSCT, therefore, the 5-year OS was 82% (95% CI, 51-96%) for autologous HSCT and 76% (95% CI, 55-90%) for allogeneic HSCT. There was no significant difference between autologous or allogeneic HSCT for EFS (p = 0.81) or OS (p = 1). Statistical analysis did not show any factors significantly associated with an impact upon EFS or OS. Excluding the 4 patients who died of TRM, the median duration from HSCT to relapse or last follow-up was 43 months (range, 4-134 months) and 56 months (range, 7-139 months) for autologous and allogeneic HSCT, respectively. The median performance score among the survivors was 100% (range, 80-100%).
This is the largest study reported to date on the outcome of children with relapsed or refractory APL. All patients were able to ultimately achieve morphologic remission with various salvage regimens and proceed to consolidative HSCT, though two patients entered HSCT positive for the PML-RARα fusion transcript. Outcomes following any form of HSCT were excellent, with a combined 5-year EFS of 72% and OS of 78%.
In the 11 children with relapsed or refractory APL (6 in molecular remission, 1 with molecular evidence of disease, and 4 unknown molecular status) undergoing autologous HSCT, we found a 5-year EFS of 73%. Reports of autologous HSCT for APL in adults have included from 10 to 50 patients each, and have demonstrated an EFS ranging from 51-90% (10-12, 23). Molecular status has not always been known prior to autologous HSCT, though de Botton, et al, demonstrated that 7-year EFS is superior for those patients in molecular remission at time of HSCT, compared to unknown molecular status (77% vs. 50%; p = 0.07) (10).
In the 21 children with relapsed or refractory APL undergoing allogeneic HSCT, we found a 5-year EFS of 71%. Roman, et al reported a group of 10 patients (including 1 child) undergoing allogeneic HSCT with an EFS of 90% (11). The only patient that relapsed was not in remission at the time of HSCT(11). Larger reports of allogeneic HSCT for APL in adults have included from 17 to 137 patients each, and have demonstrated an EFS ranging from 46-59% (10, 12, 15). The only report of allogeneic HSCT for APL in a solely pediatric population demonstrated a 5-year OS of 73% in a group of 11 patients (8 with cytogenetic confirmation) (16).
For most patients with leukemia, the primary reason for transplant failure is relapse. In children with acute myeloblastic leukemia (AML) in second complete remission, the relapse incidence following autologous HSCT ranges from 52-54% (24, 25), while the relapse incidence following allogeneic HSCT ranges from 33-42% (24, 26). In our cohort, patients undergoing autologous and allogeneic HSCT demonstrated a relapse incidence of 27% and 10%, respectively. Unfortunately, we were not able to confirm the PML-RARα status of all autograft products in this cohort. Meloni, et al confirmed the importance of a negative RT-PCR for the t(15;17) prior to autologous HSCT in their group of 15 patients (including 2 children) (14). In that series, all 7 patients with positive RT-PCR in their pre-transplant bone marrow aspirate relapsed following HSCT, while 6 of 8 patients with negative RT-PCR experienced sustained remission (14). Clearly, the ability to confirm RT-PCR negativity in both the patient and the autograft product prior to infusion may play a significant role in the successful selection of patients most likely to benefit from autologous transplant. Of note, while prior reports suggest that the median time to relapse following autologous HSCT is 5-8 months (13, 21), 2 of the 3 relapses seen following autologous HSCT in this cohort occurred over one year post-HSCT, suggesting that PCR-based screening of either peripheral blood or bone marrow for relapse should be continued on a regular basis for at least 2 years following autologous HSCT.
In the allogeneic setting, the only patients experiencing a relapse in this cohort underwent fully-matched sibling donor HSCT and did not manifest any evidence of GVHD. We postulate that the low incidence of relapse after allogeneic HSCT may be due to a high susceptibility of APL to a graft-versus-leukemia (GVL) effect. It has been shown in vitro that the PML-RARα fusion protein contains a novel antigenic site recognizable by CD4+ lymphocytes in an HLA-restricted context on antigen-presenting cells (27). Lo-Coco, et al, reported two patients with RT-PCR positive testing for disease post-allogeneic HSCT, both of whom converted to negative following withdrawal of immunosuppression, which provides evidence for an in vivo GVL effect (15). In addition, successful treatment of persistent molecular disease following MRD-HSCT with donor lymphocytes infusions has been reported (28). Other groups have reported low relapse incidences of 4%-17% following allogeneic HSCT for APL (10-12, 15), providing further support for a stronger GVL effect against residual APL as compared to non-promyelocytic AML.
In this cohort, the biggest impediment to successful outcome after allogeneic HSCT was a TRM of 19%. Possibly due to low subject numbers, we were not able to identify any predictors of TRM in patients with relapsed or refractory APL. One contributor to TRM after HSCT may be the intensity of the conditioning regimen. In patients with high-risk AML, Mengarelli, et al, reported no benefit of the addition of agents such as etoposide or cytarabine to the standard regimen of TBI plus cyclophosphamide or busulfan plus cyclophosphamide (15). In adults with APL undergoing allogeneic HSCT, several groups have shown that TBI-based conditioning is not superior to busulfan-based conditioning (10, 12). Given the low relapse rate seen in this cohort, and in other reports of HSCT for relapsed or refractory APL, it appears likely that intensified conditioning regimens will not be found to be superior to standard regimens, especially in the setting of mismatched and/or unrelated donor transplants, where the GVL effect is expected to be strongest.
Another variable that might impact TRM is the type of secondary chemotherapy regimen used to achiever remission prior to HSCT. Due to the wide variety of regimens utilized in this cohort, no regimen could be conclusively linked to an impact on HSCT-outcome. However, the importance of the pre-transplant salvage regimen has been highlighted by the case of gemtuzumab ozogamicin (GO), a monoclonal antibody against CD33, which has been shown to be highly effective in relapsed APL (15). As many as 64% of patients that receive GO within 3.5 months prior to HSCT ultimately develop evidence of sinusoidal obstruction syndrome following transplant (29). In patients with APL, several groups have shown a high incidence of TRM following intensive cytotoxic chemotherapy regimens prior to allogeneic HSCT (10, 23). As demonstrated by one patient in this cohort, successful outcomes have been reported following allogeneic HSCT despite lack of molecular remission (10, 11, 15). Some salvage regimens for APL utilize differentiating agents (such as ATRA or ATO) that lack significant cytotoxicity. Use of these agents in place of conventional chemotherapy may represent a less aggressive method by which to achieve morphologic remission prior to HSCT that could potentially translate into decreased post-HSCT TRM.
In addition to low patient numbers, the other major limitation of this study is the retrospective nature, in which we attempt to analyze the outcomes of transplants done at different centers, utilizing multiple different salvage regimens, donor types, conditioning regimens, and GVHD prophylaxis strategies. Clearly, the optimal method to analyze outcomes for patients with relapsed or refractory APL treated by HSCT would be in the context of a standardized prospective clinical trial in which inter-institutional variances could be minimized.
In summary, this study provides further evidence that for children with relapsed or refractory APL, autologous HSCT can be successfully performed for patients in remission. Other reports suggest that that molecular remission of the patient and an autologous stem cell product that is PCR negative for disease significantly decrease relapse incidence following autologous transplantation (10, 14). If molecular remission is not able to be achieved, allogeneic HSCT can be curative for many children. Few patients experience relapse of APL following allogeneic HSCT, possibly due to a strong GVL effect against residual disease. However, further improvements in supportive care are needed in order to minimize the impact of TRM in the allogeneic setting and thereby improve the overall survival of children with relapsed or refractory APL.
The authors would like to thank all of the doctors and clinical research associates who assisted in data collection. We acknowledge the ECOG and CALGB for their participation in data sharing and statistical analysis. We thank Bayard Powell, M.D. for his assistance and suggestions.
Disclosure: The authors do not have any financial or other relationships with entities that have investment, licensing, or other commercial interests in the subject matter under consideration in this article.