Abstract
Background:
Over the past 30 years, prospective randomized trials led to a markedly increased overall survival (OS) for pediatric patients with acute myeloid leukemia (AML). While hematopoietic stem cell transplantation (HSCT) was previously recommended to all patients with available matched sibling donor in first remission, recent protocols mainly restricted allogeneic HSCT to patients with very high relapse risk, refractory disease or after relapse. At the same time, the first-line treatment was heavily intensified, while supportive care was greatly improved. Also recent studies for relapsed patients demonstrated an improved, but still poor outcome (5yr-pOS: 24-38%). Thus, the factors that mainly contributed to the markedly increased OS in pediatric AML remain largely unknown. Here we present a retrospective analysis, which we performed to reveal those factors, but also to identify the obstacles that may hamper further improvement.
Patients and Methods:
From 1/1987 to 12/2012, a total of 1940 patients with de novo AML were enrolled in four consecutive trials of the AML-Berlin/Frankfurt/Muenster study group (AML-BFM-87, -93, -98, -04 and the following registry). Median follow-up was 10 years. Five year estimates of EFS and OS (pEFS, pOS), cumulative incidences (CI) of early death (ED), death in nonresponse (NR), death in complete remission (CR) or deaths of patients with one or more relapses were analyzed retrospectively over time in 3-year and 9-year periods and also according to study periods.
Results:
In this large population of 1940 patients with, we sought to systematically decipher the causes of death, incidences and survival after relapse or NR and the impact of HSCT in pediatric AML. As previously shown, the 5yr-pOS markedly increased from 52±4% (1987-1989) to 72±5% (2011-2012) (p<0.0001), but improvement of 5yr-pEFS was unsteady from 43±3% to 55±5%. This implies that factors after the first event are responsible for the better survival. Accordingly, the OS of patients with NR (n=48, OS 13±5% [AML-BFM 87], vs. n=65, OS 41±6% [AML-BFM 04]; p<0.001) and relapse (n=97, OS 19±4% [AML-BFM 87] vs. n=204, OS 42±4%, [AML-BFM 04]; p<0.001) improved remarkably (Figure 1). Treatment intensification resulted generally in less disease-related deaths in most subgroups, but in turn was associated with higher treatment-and procedure-related mortality over the last 25 years. Incidences of ED significantly decreased (p=0.002). More specifically, disease-related causes within this group decreased, whereas treatment-related deaths did not change (Figure 2). Patients with NR also showed less disease-related deaths in recent compared to the earlier cohorts (p=0.001), and no change in treatment-related lethality (p=0.43). In the recent cohort, we observed a trend of fewer deaths in CR. Treatment-related deaths decreased significantly (p=0.02). CI of death after relapse showed a remarkable decrease over time. This is mostly affecting disease-related deaths (p<0.011), while deaths due to HSCT increased (p=0.008). Deaths in patients with more than one relapse did not change, and the only significant change within this subgroup is a higher incidence of HSCT-related deaths. The increase of HSCT in first CR and after NR/relapse from 25% 1987 - 1995 to about 40% after 1995 translates into a higher incidence of procedure-treatment related deaths. Nevertheless OS of patients with HSCT after NR/relapse increased significantly (OS 35±5% [1987-1995], 37±4 [1996-2004] and 51±4% [2005-2012] (p=0.001)). The comparison within study periods showed no trend for a better pEFS and pOS in the second half of a study compared to the first so an improvement due to general measures like better supportive care cannot be proven. In a Cox regression analysis of OS there was a significant trend for better OS over the study periods (HR 0.88, 95% confidence interval 0.72-0.82) but not for better OS in the second half (HR 0.95, 0.82-1.09).
Conclusion:
In conclusion, efficacy of salvage therapy and better therapy strategy is one of the main reasons for improved outcome in pediatric AML within the last 3 decades. However, stable and in some subgroups decreased treatment-related mortality reflects better supportive care with respect to intensified therapy. Of note, procedure-related mortality increased, but nevertheless survival after HSCT improved.
OS after relapse or nonresponse
Cumulative incidences of deaths
Rasche:Jazz Pharma: Other: Travel accomodation. Reinhardt:Celgene: Research Funding; Pfizer: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees; Jazz Pharma: Other: Travel Accomodation; Boehringer Ingelheim: Membership on an entity's Board of Directors or advisory committees.
