Up-Front Randomization 10 Years Later: The AML Cooperative Group Strategy

As a global approach to acute myeloid leukemia the AMLCG started a clinical trial in 1999 with the intention to clarify the influence of baseline characteristics as well as treatment variables on patients’ outcome. The heterogeneity of this disease in mind, all randomization was done stratified for potential risk factors. Furthermore, all randomization was restricted to the single time point up-front of any treatment. Thus, an unbiased intention-to-treat analysis could be provided.

Between June 1999 and January 2012, 3370 consecutive patients 16 - 86 (median 61) years of age entered the trial at 57 participating centers. The general up-front randomization comprised

(A) repeated G-CSF priming versus no G-CSF,

(B) double induction by TAD (thioguanine-araC-daunorubicin) all at standard dose plus HAM (high-dose araC- standard dose mitoxantrone) versus HAM plus HAM,

(C) postremission TAD and maintenance versus TAD and autologous stem cell transplantation.

For age adaption patients of 60 years and older received reduced intensity double induction (B) and maintenance instead of autologous transplantation (C).

The complete remission rate was 57% (65% <60 and 51% 60+ years, p=<.0001). With a median observation time of 7.1 years the overall survival at 5 years is 26% (40% <60 years and 14% 60+ years, p<.0001) and the rate of patients in complete remission at 5 years is 29% (41% <60 years and 15% 60+ years, p<.0001). There were no significant differences in the CR rates, the 5 years survival and continuous complete remission rates according to randomizations for treatment variables.

In contrast, the CR rates and the therapeutic endpoints projected to 5 years were found significantly influenced by risk factors as identified within both age groups and including favorable cytogenetics/ molecular genetics (Döhner H et al. Blood 115: 453-74; 2010), adverse genetics, AML secondary to cytotoxic treatment or to previous myelodysplastic syndrome, LDH in serum, and in particular increasing age itself.

In patients between 16 and 59 years of age receiving equal intensity treatment those below versus above the median age of 48 years showed a CR rate of 69% versus 57% (p<.0001), a 5 years survival of 48% versus 33% (p<.0001), and a 5 years proportion in CR of 46% versus 35% (p<.0001). In patients between 60 and 86 years who received reduced intensity treatment the median age was 67 years, and those below versus above this median showed 51% versus 47% CR (p=.1), 5 years survival of 19% versus 9% (p<.001) and 19% versus 12% 5 years proportion in CR (p=.01).

Thus, up-front randomization proved to provide unbiased informations about the impact of various risk factors and treatment variables.

While age was found to dominate other risk groups (Büchner T et al. J Clin Oncol: 27: 61-69; 2009), a difference in treatment intensity by factor 3.7 had no influence of outcome in a sample of equal baseline characteristics (Büchner T et al. Blood 122: 1447; 2013). As an alternative to antileukemic cytotoxicity, allogeneic stem cell transplantation in first remission was found to prolong even the overall survival in a matched pair analysis again done prospectively within present AMLCG99 trial (Stelljes M et al. J Clin Oncol 32: 288-96; 2013).

Up-front randomization and a common standard arm also qualified as a model to compare treatment strategies between different studies (Büchner T et al. Leukemia Research 26: 1073-75; 2002, J Clin Oncol 30:3604-10; 2012). In older patients a web-based prediction of complete remission and early death could be established based upon the AMLCG99 trial (Krug U et al. Lancet 376: 2000-08; 2010).

In summary, present long-term results from up-front randomization strongly support a strict prospective strategy for cooperative clinical research on acute myeloid leukemia.