Value of Cytogenetic Abnormalities in Adult Patients with Philadelphia Chromosome (Ph)-Negative Acute Lymphoblastic Leukemia (ALL) Treated in the Pediatric-Inspired Trials from the Group for Research on Adult ALL (GRAALL)

Background: Numerous recurrent chromosomal abnormalities have been described in adult Ph-negative ALL, often observed in small patient cohorts. In the largest MRC/ECOG study (Moorman, Blood 2007), t(4;11)(q21;q23), 14q32 involvement, complex karyotype (≥5 abnormalities), and low hypodiploidy/near triploidy (Ho-Tr) were associated with shorter event-free survival (EFS), while patients with high hyperdiploidy or del(9p) had a better outcome. We aimed to confirm these observations in 955 adult patients (15-60y; median, 35y) with Ph-negative ALL treated in the pediatric-inspired GRAALL-2003/2005 trials.

Patients and Methods: Overall, a karyotype was performed for 946 (611 BCP-ALL, and 335 T-ALL), successful for 811 (523 BCP-ALL and 288 T-ALL) and abnormal in 590 patients (387 BCP-ALL and 203 T-ALL). FISH and/or PCR screening for relevant abnormalities and DNA index were also performed, finally allowing for the identification of cytogenetic abnormalities in 677/955 patients (71%). All were centrally reviewed. Ultimately, 857/955 patients (90%; 542 BCP-ALL and 315 T-ALL) could be classified in 18 exclusive primary cytogenetic subsets as detailed below. Endpoints were cumulative incidence of failure (CIF, including primary refractoriness and relapse) and EFS. With a median follow-up of 4 years, 5-year CIF and EFS were estimated in these patients at 31% and 51%, respectively. As some abnormalities, including MLL rearrangements, Ho/Tr, t(1;19)(q23;p13)/TCF3 and complex karyotypes were used to stratify allogeneic stem cell transplantation (SCT) in GRAALL trials, some comparisons were repeated after censoring patients transplanted in first CR at SCT time.

Results: The 542 informative BCP-ALL patients were classified as: t(4;11)(q21;q23)/MLL-AFF1 (n=72; 13%); other MLL+ 11q23 abnormalities (n=11; 2%); t(1;19)(q23;p13)/TCF3-PBX1 (#28; 5%); Ho/Tr (n=33; 6%); high hyperdiploidy (n=36; 7%); abnormal 14q32/IGH translocation (n=27; 5%); t(12;21)(p13;q22)/ETV6-RUNX1 (n=2; 0.4%); iAMP21 (n=3; 0.6%); other abnormalities (n=210; 39%); and no abnormality (n=120; 22%). The 315 informative T-ALL patients were classified as: t(10;14)(q24;q11)/TLX1 (n=64; 20%); other 14q11 or 7q34/TCR (n=31; 10%); t(5;14)(q35;q32)/TLX3 (n=29; 9%); t(10;11)(p12;q14)/PICALM-MLLT10 (n=14; 4%); deletion 1p32/SIL-TAL (n=18; 6%); MLL+ 11q23 abnormalities (n=6; 2%); other abnormalities (n=93; 30%); and no abnormalities (n=60; 19%). A complex karyotype was observed in 27/527 (5%) BCP-ALL and 21/298 (7%) T-ALL patients and a monosomal karyotype (as per Breems, JCO 2008) in 82/518 (16%) BCP-ALL and 26/286 (9%) T-ALL patients. In BCP-ALL, trends towards higher CIF and shorter EFS were observed in t(4;11) patients, with or without SCT censoring (HRs, 1.34 to 1.64; p values <0.10). Shorter EFS was observed in 3 subsets: 14q32 (HR, 2.10; p=0.002), Ho/Tr (HR, 1.45; p=0.10), and monosomal karyotype (HR, 1.42; p=0.029), but CIF were not different. This might be related to the older age of patients in these subsets (medians, 43y, 53y and 44y; p=0.029, <0.001 and <0.001, respectively) and worse treatment tolerance. For instance, higher incidences of non ALL-related deaths were observed in patients with 14q32 abnormalities or monosomal karyotype (p=0.031 and 0.067, respectively). Patients with high hyperdiploidy only tended to have lower CIF and longer EFS. Complex karyotype did not impact CIF and EFS, even after SCT censoring. Conversely, in T-ALL, complex karyotypes were associated with shorter EFS (HR, 2.20; p=0.004), even if the difference in CIF did not reach significance. A worse outcome was also observed in patients with t(10;11)(p12;q14)/PICALM-MLLT10 (HR, 2.45 and 2.14; p=0.016 and 0.021, for CIF and EFS respectively). A longer EFS was observed in patients with t(10;14)(q24;q11)/TLX1 (HR, 0.55; p=0.014), with a trend for lower CIF (HR, 0.59; p=0.070), while no inferior outcome was observed in t(5;14)(q35;q32)/TLX3 patients.

Conclusion: These results show that, in the context of an intensified pediatric-inspired protocol designed for adult Ph-negative ALL patients, few cytogenetic subsets remained reliably predictive of response to therapy. Differences observed in EFS might partly be due to treatment-related mortality. Combining cytogenetics, molecular genetics and minimal residual disease monitoring could allow for better individual risk assessment (Beldjord, Blood 2014).