Karyotypic Complexity In Acute Myeloid Leukemia In The Context Of Adverse Prognosis

Cytogenetic analysis is a mandatory component in the diagnostic evaluation of acute myeloid leukemia (AML) providing information regarding the biology of the disease including response or resistance to therapy. One of the cytogenetic markers which reflect an adverse outcome in conventional chemotherapy regimens is the complex aberrant karyotype consisting of multiple unrelated cytogenetic abnormalities. In AML two definitions have been established which differ in the perception of unbalanced aberrations as well as the number of single aberrations. The ELN classification scheme adopts three unrelated abnormalities while the UK MRC recently recommended four abnormalities as the most informative cut-off of complexity in the context of an adverse prognosis. The aim of this work was to study the best cut-off defining complexity (3 vs. 4) in AML with other cytogenetic high-risk markers.

The databases of three clinical multicentric, randomized, and prospective SAL trials (NCT 00180115, 00180102, and 00180167) were analyzed for AML patients with multiple cytogenetic aberrations as well as normal karyotypes (control group). Unbalanced abnormalities were counted as two aberrations according to the recommendations of the MRC (i.e. a single unbalanced translocation leading to gain and loss of chromosomal material as two unique abnormalities). The following single aberrations associated with an adverse prognosis according to ELN as well as UK MRC recommendations were included: inv(3), t(3;3), abn(3q), -5, del(5q), t(5q), t(6;9), -7, add(7q)/del(7q), t(11;v)(q23;v) (except t(9;11)), and abnl(17p).

Complete data were analyzed from 2056 patients: normal karyotype (NK) n=1590, three aberrations (K3) n=65, ≥ four aberrations (K4) n=355, t(8;21)/inv(16)/t(16;16) and at least two additional aberrations n=46. All four groups differed significantly in 5–year overall survival (OS): 35% [95% CI 32–37], 19% [95% CI 9–29], 7% [95% CI 4–10], 67% [95% CI 53–81], respectively, p≤0.001. The K4 group had a significant inferior 5–year OS as compared to the K3 group, 19% [95% CI 9–29] and 7% [95% CI 4–10], p≤0.001. HSCT was performed in first remission in 25% of patients with K3 (n=16) and 17% of patients with K4 (n=59) (p=n.s.). As demonstrated earlier, multiple aberrations additional to the good risk anomalies (t(8;21), inv(16), or t(16;16)) did not impact on the favourable prognosis of the respective group. In the K3 and K4 groups single adverse risk abnormalities were found in 55% (abnl(17p) 12%) and 83% (abnl(17p) 37%) in these patients, respectively. A hyperdiploid karyotype (HDK) with gains of whole chromosomes without any structural aberration or monosomy was present in 14% of K3 and 3% of K4-patients. Interestingly, HDK with three trisomies as well as ≥ four trisomies led to a survival similar to K4 patients without HDK. Therefore, the K3 group lost its inferior survival as compared to NK when patients with adverse risk, which induce a worse prognosis per se, as well as HDK were excluded (5y–OS: 29% [9–44] vs. 35%, [95% CI 32–37], p=n.s.). HDK patients or patients with additional single adverse risk abnormalities had a worse survival compared to NK (5y–OS: 11%, [95% CI 0–32], p=0.012; and 15%, [95% CI 3–28], p=0.004 vs. 35%, [95% CI 32–37], respectively). In contrast, when comparing the K4 group after exclusion of adverse risk and HDK patients to NK, the K4 group remained its inferior OS as compared to NK, p<0.001.

Hence, our investigation confirms and therefore favors the ≥4 cut-off of complexity in the context of an adverse prognosis as proposed by the MRC with the exception of HDK patients. HDK patients should be considered as high-risk independent of the level of complexity. Whether K3 patients without single adverse risk abnormalities and HDK should be treated as intermediate risk, as suggested by our results, needs to be investigated prospectively in clinical trials.

Platzbecker:Celgene: Honoraria, Research Funding; Novartis: Honoraria, Research Funding.