Prognostic Relevance of Cytogenetics in Childhood Acute Lymphoblastic Leukaemia (ALL): Final Results From MRC ALL97.

Abstract 88

Chromosomal abnormalities in childhood ALL are important disease markers and predictors of prognosis. Virtually all modern protocols recommend that patients with t(9;22), MLL translocations and haploidy (<30 chromosomes) receive intensive therapy. However, there is less consensus regarding the prognostic relevance of other abnormalities, e.g. t(1;19), intrachromosomal amplification of chromosome 21 (iAMP21), dic(9;20), abnormal 9p, CDKN2A deletions and low hypodiploidy (30–39 chromosomes). Furthermore, the long term prognosis and independent effect of some abnormalities, especially t(12;21)/ETV6-RUNX1, has been questioned.

We investigated the prognostic relevance of cytogenetics among 1,934 children treated on MRC ALL97. Patients with t(9;22), haploidy, low hypodiploidy and those aged <2yrs with a MLL translocation were treated as high risk. In order to focus on the intrinsic aggressiveness of the leukemic clone, we used relapse-free survival (RFS) as our primary endpoint. In addition, we constructed a cytogenetic based risk index to assess the utility of cytogenetics as a whole in predicting relapse.

The 5yr RFS of the whole cohort was 81% (95% CI 79–82%) with a median follow-up of 8.2yrs. Univariate analysis revealed 5 abnormalities that were significantly associated with relapse: t(12;21) Hazard ratio (HR)=0.50 (95% CI 0.36, 0.68); high hyperdiploidy (51–65 chromosomes): 0.58 (0.45, 0.74); iAMP21: 5.51 (3.57, 8.50); t(9;22): 3.31 (2.06, 5.32); other MLL translocations [not t(4;11)]: 2.70 (1.39, 5.25); and 17p loss [del(17p)]: 2.13 (1.35, 3.34) (all p<0.003). Moreover, all abnormalities, except other MLL translocations, retained their significance in multivariate analysis. The following abnormalities were not predictive of relapse: t(4;11), t(1;19), dic(9;20), CDKN2A deletions, −7 and abnormal 9p. There were too few haploid and low hypodiploid patients to be formally tested but 10/18 (56%) relapsed.

Further analysis of high hyperdiploid patients revealed that there was no difference in RFS for those with (n=218) and without (n=200) triple trisomy (+4, +10 and +17) [HR=0.81 (0.49, 1.34), p=0.4]. However, high hyperdiploid patients with a +18 (n=396) had a lower risk of relapse [HR=0.44 (0.26, 0.74) (p=0.002)] than other patients (n=86).

Among 369 t(12;21) patients 47 (13%) suffered a relapse. The timing of these relapses was different to the rest of the cohort with fewer early relapses (within 6 months of the end of treatment) in the t(12;21) cohort: 9/47 (19%) versus 170/285 (59%), p<0.001. The 5yr and 7yr cumulative risk of relapse for t(12;21) patients was 11% (95% CI 8–15%) and 13% (9–17%) respectively compared to 24% (22–28%) and 27% (24–30%), respectively for other patients.

A total of 54/1596 (3.4%) patients had del(17p) by cytogenetics: i(17q) (n=18), del(17)(p) (n=7), monosomy 17 (n=8) and unbalanced translocations (n=21). It is a secondary abnormality co-existing with high hyperdiploidy (n=21), t(12;21) (n=6), MLL translocations (n=2) and t(9;22) (n=1). Overall, these patients had an inferior RFS (64%, (49%–75%), p=0.004). However, the presence of del(17p) did not abrogate the good outcome of patients with t(12;21) and high hyperdiploidy [HR=1.70 (0.75, 3.87) p=0.206] but did represent a strong marker of relapse among other patients [HR=2.96 (1.68, 5.20) p<0.001].

We classified 1,733 patients into three cytogenetic risk groups: good (GRG) [t(12;21), high hyperdiploidy] n=928 (54%); poor (PRG) [t(9;22), t(4;11), other MLL, t(17;19), haploidy, low hypodiploidy, iAMP21, del(17p)] n=153 (9%); standard (SRG) [all other cases] n=652 (38%). The 5 year RFS were GRG 88% (85–90%), SRG 79% (75–82%) and PRG 49% (40–57%). In multivariate analysis, patients in the GRG or PRG were less or more likely to relapse compared to patients in the SRG: HR=0.65 (0.50–0.83), p=0.001 and HR=2.67 (2.01–3.53), p<0.001, respectively. Moreover, there was a strong correlation between cytogenetic risk group and the BFM risk classification of relapses, based on immunophenotype, timing and site of relapse. Among GRG patients who relapsed, only 11/130 (8.5%) suffered a high risk relapse; whereas 36/76 (47%) PRG patients who relapsed had a high risk relapse.

These data clarify the prognostic relevance of several chromosomal abnormalities in the context of a modern childhood ALL therapy and provide further evidence that cytogenetics is a powerful and independent indicator of relapse risk.