Prognostic Utility of the European LeukemiaNet (ELN) Genetic-Risk Classification in Adults with De Novo Acute Myeloid Leukemia (AML): A Study of 1,550 Patients (Pts)

Mrózek, Krzysztof; Marcucci, Guido; Maharry, Kati; Nicolet, Deedra; Becker, Heiko; Whitman, Susan P.; Metzeler, Klaus H.; Schwind, Sebastian; Wu, Yue-Zhong; Kohlschmidt, Jessica; Pettenati, Mark J.; Koduru, Prasad R. K.; Heerema, Nyla A.; Block, Annemarie W.; Patil, Shivanand R.; Baer, Maria R.; Kolitz, Jonathan E.; Moore, Joseph O.; Carroll, Andrew J.; Larson, Richard A.; Bloomfield, Clara D.

doi:10.1182/blood.V118.21.414.414

Abstract

Abstract 414

The ELN standardized system for reporting cytogenetic and select molecular findings was created to enable comparisons of studies correlating treatment outcome with genetic data in AML (Döhner et al. Blood 2010;115:453). The ELN classification is based on published reports on prognostic import of cytogenetic and molecular alterations, but has not been tested in large cohorts of similarly treated AML pts, except for 1 study comprising ∼20% secondary AML pts (Röllig et al. JCO 2011;29:2758). Thus our goal was to assess the prognostic utility of the ELN classification in a relatively large cohort lacking the confounding effects of AML type (de novo v secondary) or different therapies [chemotherapy v allogeneic stem cell transplant (SCT)]. We analyzed 1,550 de novo AML pts (excluding AML M3, which is not part of the ELN classification) treated with cytarabine-daunorubicin-based chemotherapy on Cancer and Leukemia Group B frontline protocols; no pt received SCT in 1^st complete remission (CR). Since consolidation therapy of younger [<60 years (y)] and older ('60 y) pts differed in intensity, we analyzed younger (n=818; median age 44 y, range 17–59) and older (n=732, median age 69 y, range 60–86) pts separately.

The pts were classified into the 4 ELN Genetic Groups: Favorable (Fav), Intermediate I (Int I), Intermediate II (Int II) or Adverse (Adv). Younger pts were classified in the Fav Group twice as often and in the Int II and Adv Groups ∼30% less often than older pts (Table 1). Within the Fav Group, core binding factor AML [t(8;21), inv(16) or t(16;16)] was twice as common (P<.001) and NPM1-mut/FLT3-ITD^– half as common in younger than in older pts. In the Int I Group, NPM1-mut/FLT3-ITD⁺ was more common in younger and NPM1-wt/FLT3-ITD^– in older pts, whereas in the Int II Group, t(9;11) was twice more common in younger pts. In the Adv Group, pts with balanced cytogenetic abnormalities (abns), ie, inv(3) or t(3;3), t(6;9) and t(v;11)(v;q23), were mainly younger, and pts with –7 mainly older (Table 1).

Table 1.

Distribution of ELN Groups and subsets in 818 younger and 732 older pts

Genetic Group Subset	Younger Pts %*	Older Pts %*	P
Fav	41	20	<.001
t(8;21)	22	13	.01
inv(16) or t(16;16)	33	12	<.001
NPM1-mut, no FLT3-ITD^†	28	54	<.001
CEBPA-mut^†	17	20	.44
Int I	18	19	.64
NPM1-mut & FLT3-ITD^†	55	38	.006
NPM1-wt & FLT3-ITD^†	10	12	.85
NPM1-wt, no FLT3-ITD^†	35	50	.01
Int II	19	30	<.001
t(9;11)	11	5	.05
Cytogenetic abns not classified as Fav or Adv	89	95	.05
Adv	22	31	<.001
inv(3) or t(3;3)	8	3	.03
t(6;9)	4	1	.02
t(v;11)(v;q23)	15	6	.01
−5 or del(5q)	3	3	.78
−7	5	12	.02
abn(17p)	0	0	—
Complex karyotype (≥3 abns)	65	75	.04

Genetic Group Subset	Younger Pts %*	Older Pts %*	P
Fav	41	20	<.001
t(8;21)	22	13	.01
inv(16) or t(16;16)	33	12	<.001
NPM1-mut, no FLT3-ITD^†	28	54	<.001
CEBPA-mut^†	17	20	.44
Int I	18	19	.64
NPM1-mut & FLT3-ITD^†	55	38	.006
NPM1-wt & FLT3-ITD^†	10	12	.85
NPM1-wt, no FLT3-ITD^†	35	50	.01
Int II	19	30	<.001
t(9;11)	11	5	.05
Cytogenetic abns not classified as Fav or Adv	89	95	.05
Adv	22	31	<.001
inv(3) or t(3;3)	8	3	.03
t(6;9)	4	1	.02
t(v;11)(v;q23)	15	6	.01
−5 or del(5q)	3	3	.78
−7	5	12	.02
abn(17p)	0	0	—
Complex karyotype (≥3 abns)	65	75	.04

*

For genetic subsets, % are calculated within each Genetic Group

^†

Pts with cytogenetically normal AML

After a median follow-up for younger and older pts alive of, respectively, 7.6 y (range, 0.6–19.1) and 6.1 y (range, 2.3–16.4), CR rates, disease-free (DFS) and overall (OS) survival differed significantly across the ELN Groups in both age groups (Table 2, Fig. 1). Pairwise comparisons between individual ELN Groups also yielded significant differences for all of them, except for the comparison between the Int I and Int II Groups, which did not show significant differences for any outcome endpoint in older pts nor for CR rates or DFS in younger pts; OS of younger Int II pts was longer than OS of Int I pts (P=.02; Fig. 2). For each ELN Group, the outcome of older pts was worse than that of younger pts. We conclude that despite differences in frequencies of younger and older pts classified into Fav, Int II and Adv ELN Groups and subsets within these Groups, the ELN classification allows prognostic grouping of both younger and older de novo AML pts. It is hoped that new markers (eg, TET2, ASXL1 mut) will improve pt classification to ELN Int I and Fav Groups.

Table 2.

Outcomes according to the ELN Groups

Outcome	Fav	Int I	Int II	Adv	P
Younger pts, n=818	n=339	n=144	n=156	n=179
CR rate, %	96	76	79	50	<.001
DFS, median, y	5.5	0.8	1.2	0.6	.001
OS, median, y	11.5	1.2	2.1	0.8	<.001
Older pts, n=732	n=145	n=136	n=222	n=229
CR rate, %	83	61	63	39	<.001
DFS, median, y	1.1	0.6	0.7	0.5	<.001
OS, median, y	1.6	0.9	0.9	0.5	<.001

Outcome	Fav	Int I	Int II	Adv	P
Younger pts, n=818	n=339	n=144	n=156	n=179
CR rate, %	96	76	79	50	<.001
DFS, median, y	5.5	0.8	1.2	0.6	.001
OS, median, y	11.5	1.2	2.1	0.8	<.001
Older pts, n=732	n=145	n=136	n=222	n=229
CR rate, %	83	61	63	39	<.001
DFS, median, y	1.1	0.6	0.7	0.5	<.001
OS, median, y	1.6	0.9	0.9	0.5	<.001