DNX added to FLAG improves early treatment response in pediatric relapsed AML. Overall long-term survival was similar, but CBF-AML showed an improved survival with FLAG/DNX. International collaboration proved feasible and resulted in the best outcome for pediatric relapsed AML reported thus far.

No abstract

This study shows that single-agent MTX in the applied regimen in pediatric relapsed AML has limited efficacy. However, it also demonstrates the feasibility of an international and therapeutic window phase II study in pediatric relapsed AML.

18 Introduction: Long-term survival in pediatric relapsed AML is only 20-30%. Optimal reinduction therapy is unknown, and there is a concern about cardiotoxicity with repeated anthracycline use at relapse. Preclinical in vitro and animal studies, and limited clinical data suggest that liposomal daunorubicin (DaunoXome®, DNX) is less cardiotoxic. These considerations lead to a phase III study, in the setting of the International BFM Study Group. Materials and methods: FLAG was randomised against FLAG/DNX in the 1st reinduction course. The conventional 5-days FLAG only was recommended as the 2nd course. DNX was dosed at 60 mg/m2/day on days 1, 3 and 5. After induction, allogeneic stem cell transplantation was generally recommended, but time-to-transplant could be bridged by high- or low-intensive consolidation therapy. Primary endpoint of the study was early treatment response, based on bone marrow examination shortly before reinduction course 2, and defined as either good (≤20% leukemic blasts) or poor (>20% leukemic blasts). This endpoint was chosen because of its prognostic value in earlier relapsed AML BFM-trials, and because compliance with an extended protocol guideline was likely to be suboptimal within the context of a highly multinational and multicenter AML Relapse protocol. However, secondary endpoints were defined, including the CR2 rate determined after 2 courses, long-term survival, and toxicity. Patients with AML M3 and those >18 years of age at initial diagnosis were ineligible. The study opened in most countries in 2002/2003. The study closed for accrual on April 1, 2009 when the required 360 fully eligible and evaluable patients had been randomized. Early and late relapsed AML was defined as a relapse within or after 1 year from initial diagnosis, but this only influenced treatment in that early relapsed AML patients were eligible for haploidentical SCT, while late relapsed AML patients were eligible for autologous SCT, if a matched or partly mismatched transplant was not possible. Thirteen groups from 20 countries and >100 centers have enrolled patients, with informed consent and after approval of the study by regulatory authorities. Data are presented according to intention-to-treat, with a median follow-up of 2.7 years for patients at risk. Results: Overall 4-year probability of survival (pOS) was 35% SE 2%, the overall CR2 rate 62%. The good early responders had a 4-year pOS of 45% SE 3% versus 10% SE 3% for poor responders (p<0.0001). Similar differences were seen in the subgroups early relapse (pOS 36% SE 5% vs 5% SE3%) and late relapse (53% SE 4% vs 17% SE7%). For the main endpoint, patients randomized to FLAG/DNX had a 12% higher early good response rate than patients randomized to FLAG, 81% vs 69% (p= 0.009). Within early relapsed AML (52% of all patients), this difference was 17% in favor of DNX (p=0.025), in late relapsed AML it was 8% in favor of DNX (p=0.12). In terms of secondary endpoints, CR2 rate was 10% higher with DNX (68% vs 58%, p=0.047). Overall survival was also higher with DNX, but not significantly: 4-years pOS FLAG 35% SE 4% vs. 40% SE 4% for FLAG/DNX. Toxicity was similar in both groups and was significantly different only for skin eruptions (FLAG 1%, FLAG/DNX 5%, p=0.04). Grade III/IV cardiotoxicity was seen in 5 (3.2%) patients treated with FLAG/DNX (1 grade IV, 4 grade III), and in 2 (1.4%) patients treated with FLAG (1 grade IV, 1 grade III), often coinciding with severe infections and was treatable. Conclusion: adding DNX to FLAG improves treatment response in pediatric relapsed AML, both in terms of BM status after the 1st reinduction course and the CR2 rate after 2 courses. The 5% improved pOS with DNX and FLAG was not statistically significant, but the study was not powered to detect small differences in survival. Furthermore, short-term toxicity of FLAG/DNX was similar to FLAG. Analysis of long-term toxicity will require longer follow-up. In view of these data, standard reinduction chemotherapy for pediatric relapsed AML according to I-BFM-SG protocols should consist of high-dose ara-C plus DNX, but the cumulative dose of anthracyclines should always be taken into account. In addition, pediatric relapsed AML is curable in a subgroup of patients, particularly those with a good early treatment response, irrespective of whether the relapse of AML was early or late. Finally, international collaboration proved to be feasible in the treatment of pediatric relapsed AML, and resulted in the best outcome reported sofar with pOS of 35%. Disclosures: Off Label Use: Liposomal daunorubicin (DaunoXome®) in pediatric relapsed acute myeloid leukemia.

No abstract

184 Acute Myeloid Leukemia - Therapy, excluding Transplantation: Pediatric and Adult AML Therapy Introduction: With improving initial antileukemic therapy, CNS disease might become more important in AML. We therefore evaluated the incidence of CNS involvement in a large series of children with first relapsed AML. In addition, clinical and biological features of children with and without CNS involvement at relapse were compared, and finally the prognostic significance of CNS involvement at relapse was studied. Materials and Methods: Patients were selected from those registered in the setting of study Relapsed AML 2001/01 (ISRCTN code 94206677), based on having first relapsed AML and precise information on the location of the relapse, and on the type of CNS involvement. The latter was distinguished in asymptomatic and symptomatic disease. The protocol prescribed intensive intrathecal triple chemotherapy in case of CNS disease: first dose immediately before start of reinduction course 1. Second and subsequent doses every 7 days until 1 week after complete blast clearance of the CSF or disappearance of radiological abnormalities. Then, 2 more doses were given, one immediately before the start of the second reinduction course, and the other at the start of consolidation treatment. Cranial radiotherapy was not generally recommended and was actually used in 18% of patients with CNS relapse. Systemic therapy consisted of FLAG with or without liposomal daunorubicin (1:1 randomisation), followed by FLAG and allogeneic stem cell transplantation. This clinical study also enrolled patients with a combined relapse, or an isolated extramedullary relapse, or a bone marow relapse (isolated or combined) with <20% blasts in the BM, and these patients were also eligible for randomisation. The study was closed for enrollment in April 2009, and median follow-up for patients at risk is well above 3 years. Results: Out of 477 patients, 45 (9.4%) had CNS relapse, which percentage did not differ significantly between individual study groups. The far majority (41/45) of these patients had a combined relapse, only 4 patients had isolated CNS relapse. All 4 isolated relapses concerned early relapse, as compared to 66% of patients with a combined relapse including the CNS and to 48% of patients with first relapsed AML not involving the CNS (p-trend = 0.004). For further analyses, isolated and combined relapses involving the CNS were combined. Patients with CNS relapse were younger at relapse, mean 5.6 (SD 5.2) vs 9.7 (SD 5.3) years, p<0.0001. There was a trend for a higher percentage of males among those with CNS involvement, 71 vs 57%, p=0.07, and a trend for a higher WBC (mean 22.1 vs 13.4 × 109/l, p=0.09). Patients with CNS relapse more often had AML FAB type M5 (48 vs 16%, p<0.001), and more often MLL gene rearrangements (26 vs 11%, p=0.049) than patients with non-CNS relapse. Clinical outcome did not differ significantly between both groups, and patients with relapsed AML and CNS involvement had a 58% CR2 rate (vs 64% for the remaining patients), and a 4-years pOS of 32% (SE 8%) vs 37% (SE 2%) for the remaining patients. The majority of events within patients with relapsed AML and CNS involvement concerned refractory disease (12/45 patients), while early death occurred in 7 patients, death in CR in 3 patients, and 9 patients developed a 2nd relapse (in 3/9 patients again involving the CNS). Relapsed patients with CNS involvement had a higher rate of second complete remission when treated with liposomal daunorubicin, than if treated with FLAG only (82 v 38%, p=0.019), but overall survival was not significantly different between both treatment arms in this subgroup of patients. Conclusion: Isolated CNS relapse is rare in pediatric AML, but CNS involvement is observed in more than 9% of patients with first relapsed AML. CNS involvement is more frequent in younger patients, males and those with a high WBC at relapse. Biologically, patients with CNS involvement at relapse more frequently have acute monoblastic leukemia and MLL gene rearrangement, than patients with non-CNS relapse. Despite CNS disease at relapse, second complete remission was achieved in more than half of patients, and long-term survival was achieved in about one third of patients. Thus, intensive salvage treatment of pediatric patients with relapsed AML and CNS involvement is justified. Prevention of CNS relapse by improved therapy for newly diagnosed AML is warranted, with special attention to several subgroups which seem at higher risk for CNS relapse. Disclosures: Off Label Use: Liposomal daunorubicin in pediatric relapsed AML.

Relapse occurs in 30–40% of newly diagnosed AML patients, with long-term survival in 20% of them. We initiated a study for relapsed AML excluding AML M3 and those &gt;18 years of age. As backbone, FLAG is being used for 2 consecutive courses: fludarabine 30 mg/m2/day x 5, cytarabine 2 g/m2/day x 5, G-CSF 200 μg/m2/dose for 6 days, starting day −1. Liposomal daunorubicin (DaunoXome®, DNX) is a “new” anthracycline with potentially less cardiotoxicity, applicable in the total group of patients, and with the perspective to be useful at initial treatment as well. Therefore, DNX at 60 mg/m2/day on days 1, 3 and 5 was added or not in a 1:1 randomised fashion to the first course of FLAG. The ongoing prospective study must determine the efficacy and toxicity of DNX when added to FLAG, and the long-term outcome in a large group of relapsed AML patients. 13 groups worldwide are enrolling patients. 303 patients were registered as per April 2005. This planned interim analysis with blinded efficacy data concerns 226 eligible and evaluable patients with first relapsed AML, 140 (62%) being male. Median age at relapse was 9.7 (range, 0.7–18) years. 50% of patients relapsed within 1 year from initial diagnosis. The majority concerned isolated bone marrow relapse (85%), with 1% isolated and 5% combined central nervous system involvement. Median WBC at relapse was 4.0 (range, 0.4–293) x 109/l. Dominating FAB types are M2 with auer rods, M4 without eosinophils and M5. The randomisation was actually performed in 81% of patients. A total of 63% of patients was transplanted, the majority with a matched unrelated donor. Poor response to the 1st course of therapy (&gt;20% of blasts in the BM shortly before the 2nd course), was seen in 23% of patients. Early death occurred in 6% of patients. Complete remission (CR) was achieved in 62% of patients, and they have a probability of survival at 2 years (2-yr pSurv.) of 47±6% (compared to 33±5% for the total group), 38±9% for early relapses and 54±7% for late relapses. Patients not achieving CR had a 2-yr pSurv. of 5±4%. Compared to early relapses, patients with late relapse had higher CR rates (74 vs 51%), and higher 2-yr pSurv.: 42±6% vs 21±5%. There was significant grade III/IV toxicity, but no unexpected toxicity, and no clinically relevant differences between the arms with and without DNX, especially not in cardiotoxicity. In conclusion, it is feasible to perform a large randomised study in a very international setting. DNX added to FLAG does not result in major additional toxicity, but follow-up of cardiotoxicity should be extended as planned in this protocol. Late relapses do better in terms of CR and overall survival, but early relapses achieving CR have a realistic chance of survival as well. The study is ongoing until 360 patients have been randomised, to answer the question whether liposomal daunorubicin improves outcome in pediatric relapsed AML. Meanwhile, overall outcome is encouraging, with CR achieved in 62% of patients of whom 47% are survivors at 2 years.

Relapse remains the commonest adverse event in newly diagnosed AML patients, and the reported long-term survival after relapse in pediatric AML is 20–30%. Allogeneic stem cell transplantation (allo-SCT) in CR1 is employed by several groups with the aim of reducing the relapse rate. There is a widely held view that patients with AML who relapse after an allo-SCT in CR1 have no or minimal chances of long-term survival. However, such a cohort has never been described in the literature. Therefore, the outcome of children with AML who received an allo-SCT in CR1 and who subsequently relapsed and were enrolled in study Relapsed AML 2001/01 is described. This is a prospective, randomised study for relapsed AML, excluding AML M3 and those &gt;18 years of age at initial diagnosis comparing FLAG with Daunoxome-FLAG. FLAG is given for 2 consecutive courses: fludarabine 30 mg/m2/day × 5 days, cytarabine 2 g/m2/day × 5 days, G-CSF 200 μg/m2/dose × 6 days, starting day −1. Liposomal daunorubicin (DaunoXome®, DNX; a new anthracycline with potentially less cardiotoxicity) at 60 mg/m2/day on days 1, 3 and 5 is added or not in a 1:1 randomised fashion to the first course of FLAG. After two courses, patients are eligible for allo-SCT, sometimes bridged by high-or low-intensity consolidation chemotherapy. Efficacy data for both arms are still blinded as the study remains open until the end of 2008. Thirteen groups worldwide are enrolling patients. More than 500 patients were registered by April 2008, but this analysis relates to 329 fully evaluable patients diagnosed before July 2007, who are known to have or have not received an allo-SCT in CR1. Reasons for allo-SCT in CR1 differ between groups, but in general it is reserved for higher-risk patients and predominantly those with a matched-sibling donor. As part of relapsed treatment, 214 patients underwent allo-SCT to consolidate second line therapy, and for 11 patients this was their second allo-transplant. Twenty-two (7%) of these 329 patients had undergone allo-SCT in CR1, 12 boys and 10 girls, the majority being between 1 and 10 years of age. Three out of these 22 patients had a favorable karyotype, i.e. inv(16) or t(8;21). The majority of these patients (17/22, 77%) had a late relapse (occurring at least one year from initial diagnosis) compared to 47% of relapsed AML patients who had not received an allo-SCT in CR1 (p=0.01). There was no significant difference in chemotherapy related morbidity or mortality between those children who had received an allo-SCT in CR1 and those who had not, including cardiotoxicity. The percentage of so-called poor early responders (more than 20% bone marrow blasts on day “28” sampled shortly before the 2nd reinduction course) was 24% versus 23% (n.s.) and the CR2 rate 45% versus 65% (p=0.09) in the yes and no allo-SCT in CR1 patients, respectively. Within late relapsed AML only, the rate of poor early responders was 23% versus 15% (n.s.) and the CR2 rate 50% versus 77% (p=0.03) in the yes and no allo-SCT in CR1 patients, respectively. Finally, 4-year probability of survival from relapsed AML was 27% (95%-C.I. 8–47%) in the “yes allo-SCT in CR1” subgroup compared to 33% (27–39%) in the remaining patients (n.s.). Limiting this analysis to late relapsed AML patients, it was 29% (6–53%) versus 45% (36–54), respectively (n.s.). Five of the 6 patients who survived after allo-SCT in CR1 (median follow up 2 years) received a second allo-SCT. In conclusion, the small number of trial patients who relapsed early and who had received an allo-SCT in CR1 suggests that clinicians may have been reluctant to enter these patients into trial. In general, patients with an early relapse of AML treated in this study have a reasonable chance of cure (Kaspers et al., ASH 2007 and 2008), but the very small number of such patients who had an allo-SCT in CR1 preclude any firm conclusions for that subgroup. Patients with a late relapse who have received an allo-SCT in CR1 appear to have a somewhat (but statistically not significant) inferior outcome compared to patients who were not transplanted in CR1. The probability of survival at 4 years from relapse of nearly 30% warrants the use of second line therapy in patients with a late relapse despite allo-SCT in CR1.