SNP6 Array Better Defines Chronic Lymphocytic Leukemia (CLL) Prognostic Groups

Abstract 3611

CLL has a very heterogenous clinical course. To identify genomic lesions associated with a poor outcome, we have applied a high-density genome-wide DNA profiling on a large series of CLL cases, divided into test and validation series.

327 CLL samples collected at diagnosis from four Centers were analyzed with the Affymetrix Human Mapping SNP6.0 arrays. CLL diagnosis and management were based on the NCI Working group criteria (Hallek et al., Blood 2008). All cases were used to define the minimal common regions (MCR). TP53 inactivation was defined as the presence of del(17p) by FISH or by SNP6 or by the presence of TP53 somatic mutations.

In order to evaluate the impact of the 436 identified MCR on OS, univariate analysis was performed in a test mono-institutional series of 148 CLL cases with the log-rank test followed by the multiple test correction to control false positive rate (q-value). Twenty-seven MCRs bearing a p-value < 0.05 and a q-value < 0.1 were evaluated in a multi-institutional validation series of 179 consecutive CLL cases. The training and validation series were comparable in terms of Binet stage, IGHV mutational status, FISH prognostic groups, TP53 mutational status, telomere length, serum B2MG and LDH. Ten MCRs presented a P <0.05 also in the validation series: gains of the short arm of chromosome 2 [2p, 2p25.3-p22.3 (MYCN), 2p22.3, and 2p22.1-p14 (REL)], gains at 8q23.3-q24.3, losses at 17 (whole 17p, 17p13.3-p11.2, 17p12-p11.2), del(8p) and 9p21.3 (CDKN2A). Lesions on 2p, 8p, and 8q were validated by FISH. The 9p21 deletion, homozygous, occurred in only two cases.

All the 327 patients were pooled to improve the power of further analyses regarding the impact of these parameters on the clinical course. 8q23.3-q24.3 gains were associated with IGHV translocations (37% v 9%; P 0.007), short telomeres (75% v % 33%; P < 0.016) and, inversely, with ZAP70 (0% v 38%; P 0.038). Del(8p) was associated with IGHV3-23 (40% v 9%; P 0.001). Patients with 2p gains more frequently had U-IGHV (85% v 35%; P < 0.001), IGHV1-69 (46% v 9%; P < 0.001), ATM loss by FISH (23% v 7%; P 0.027), IGHV translocations (33% v 9%; P 0.014), short telomeres (100% v 46%; P < 0.001), increased LDH (50% v 21%; P 0.030), and transformation to Richter's syndrome (22% v 5%; P 0.021).

Del(8p) was associated with del(17p13-p11) and with 8q23.3-q24.3 gains. The latter were more frequent in cases with del(17p), del(8p) and 3q and 17q gains. All but one case bearing 2p gains had concomitant del(13q14.3), as also shown by diagnostic FISH (77% v 48%; P 0.027). Other concomitant lesions were losses at 20p, 18p, 6q21. Del(17p) occurred in only 3/13 with +2p, still higher than in the remaining cases (P 0.036).

Del(8p), gains at 2p or at 8q better defined prognostic groups. Del(8p) or 8q gains split cases with Binet stage A, older than 60, with an inactive TP53, M-IGHV, U-IGHV. Gains at 2p split patients with Binet stage A, younger or older than 60, with an apparently active TP53, M-IGHV or U-IGHV.

Cases with concomitant del(8p) and TP53 disruption by mutation and/or deletion showed poorer outcome than patients with single aberrations. Similarly, patients with +2p/del(8p)/+8q23.3-q24.3 had a shorter OS than patients carrying the individual lesions. The presence of only one, or of both lesions did not apparently affect the outcome in the following situations: del(8p)/+8q23.3-q24.3, +8q23.3-q24.3/TP53 inactivation, +2p/TP53 inactivation.

In a cox regression model adjusted for age, gender, Binet stage, and IGHV mutational status, TP53 inactivation, gains at 2p (H.R. 2.2; 95% CI, 1.1–4.5) and 8q (H.R. 3; 95% CI, 1.3–7) gains maintained their prognostic significance on OS.

In conclusion, the addition of probes targeting 2p, 8q and 8p or of a genome-wide arrayCGH to the standard FISH panel would better refine CLL prognostic groups.