Abstract
Philadelphia chromosome positive (Ph+) acute lymphoblastic leukemia (ALL) is traditionally considered the subtype with the worst prognosis, despite recent improvements in long-term survival brought about by the use of tyrosine kinase inhibitors (TKI) such as imatinib or dasatinib. Allogeneic stem cell transplantation (aSCT) remains the most effective curative post-remission therapy in adults but appears to be less critical in children, indicating a substantial clinical and biological heterogeneity within the subgroup of Ph+ ALL. The ability to segregate Ph+ ALL into subgroups with different prognosis on the basis of reductions of BCR-ABL1 transcript levels during therapy lends further support to the heterogeneity of this type of leukemia, for which the genetic basis is not known. Microarray-based genome-wide profiling studies conducted predominantly in pediatric ALL patients have recently revealed novel recurrent submicroscopic aberrations of genes involved in B-cell development and cell cycle regulation, such as CDKN2A/B, IKZF1, PAX5, ETV6, RB1, BTG1 and EBF1. Deletions of IKZF1, CDKN2A/B and PAX genes have received the most attention due to their high frequency particularly in BCR-ABL1-positive ALL and their association with an inferior prognosis in the setting of combined TKI and chemotherapy. Their prognostic relevance in the setting of allogeneic SCT for adult or pediatric high risk BCP-ALL is not known. We therefore examined whether the negative prognostic role of IKZF1 aberrations and other frequent microdeletions of genes associated with B-cell development can be overcome by allogeneic SCT in CR1.
A total of 137 newly diagnosed Ph+ ALL pts. (median age 42 years, range 18-64y, 79 male 58 female) treated within the prospective multicenter GMALL study 07/03 were analyzed. 96 of these patients underwent aSCT in first complete remission (CR), 8 pts. were primary refractory, 12 CR pts. did not undergo aSCT and relapsed, 11 pts. died during induction. Genome-wide copy number analysis in search for acquired copy number alterations (CNA) was performed with Affymetrix SNP 6.0 arrays with anonymous references. Copy number polymorphisms were excluded from the data by comparison with known copy number polymorphisms registered in the UCSC genome browser http://genome.ucsc.edu/, (hg-18). Putatively acquired CNAs were validated by multiplex ligation-dependent probe amplification (MLPA) and germline matched SNP array analysis of n=20 samples within the study.
Of the 96 pts. transplanted in CR1, 48 remain in CR (CCR), 30 pts. relapsed after aSCT and 7 died of treatment related causes, survival data only are available for one patient. CDKN2A/B genomic alterations were identified in 41% (40/97) of patients, deletions of IKZF1 and PAX5 were observed in 61% (59/97) and 39% (38/97) of pts., respectively.
Univariate analysis of the complete cohort revealed that deletion of CDKN2A/B was the only aberration with a statistically significant negative effect on overall survival (OS) (p=0.003). Among patients transplanted in CR1, IKZF1-deletions were associated with inferior median time to relapse after SCT (56 mos vs. n.r., p=0.01), DFS from SCT (15.6 mos. vs. n.r.; p=0.024) and OS (median 40 mos. vs. not reached (n.r.) p=0.04) compared with the IKZF1 wildtype cohort. Similarly, the prognosis of pts. with CDKN2A/B deletions was inferior in terms of DFS (median 10.6 mos. vs. n.r.; p=0.022) and OS (median 25 mos. vs. n.r.; p=0.01), but not of remission duration from SCT. PAX5 (p=0.07) but not the combination of all three lesions (p=0.14) showed a trend to a worse prognosis. Of the more uncommon genetic aberrations BTLA, EBF1, ETV6, RB1 and BTG1, only the latter was associated with a lower probability of remaining in CR (0% vs. 67% at 5 years; p=0.012) or DFS (0% vs. 52% at 5 years; p=0.043), with a trend towards shorter OS (median 35 mos. vs. 87 mos; p=0.078).
In conclusion, genomic lesions of IKZF1, CDKN2 and PAX5 identify a subgroup of Ph+ ALL pts. who have an inferior survival despite undergoing aSCT in CR1. Their poor outcome is attributable primarily to a high relapse rate after SCT, emphasizing the need to introduce additional treatment elements prior to and after aSCT.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.