Late Relapses of Childhood T-ALL Are Frequently Second T-ALL.

The vast majority of relapses in childhood T-cell acute lymphoblastic leukemia (T-ALL) patients occur relatively early, usually within 2 years from diagnosis. Our previous comparative molecular analyses between diagnosis and relapse in all such “classical” T-ALL showed totally or at least partly identical T-cell receptor (TCR) gene rearrangement patterns at both disease phases. These results confirm that the relapse clone in these patients originated from the original diagnosis clone, which became resistant to the applied treatment. In contrast to these “classical” T-ALL, two patients experienced very late T-ALL recurrences and displayed completely different TCR gene rearrangement sequences between diagnosis and relapse. We hypothesized that such late “relapses” of T-ALL in fact might represent second malignancies in genetically predisposed persons. We investigated 16 T-ALL patients with late relapses, i.e. at least 2.5 years from initial diagnosis. The studies at the DNA level involved detailed comparison of TCR gene rearrangements between diagnosis and relapse (PCR-heteroduplex, sequencing and/or Southern blot analyses), the detection of gene fusions involving the TAL1 gene and/or TCR genes and comparative genomic hybridization (CGH) using high-resolution Agilent arrays. We found evidence of a common clonal origin between diagnosis and relapse in ten of the 16 patients. In one case, no clonal TCR rearrangements were detected neither at diagnosis nor at relapse. In the other five patients TCR gene rearrangement sequences had completely changed between diagnosis and relapse, suggesting a second T-ALL rather than a relapse. Moreover, three of these five patients remained in complete remission after second-line treatment, which is unusual for relapsed T-ALL. In three of the five patients with presumed second T-ALL, CGH arrays showed completely different patterns of genomic aberrations between diagnosis and relapse, while in the remaining two patients the patterns of genomic changes were vastly different, but some aberrations were similar at both disease phases. In eight patients with evidence of a common clonal origin between diagnosis and relapse at least 50% of genomic events revealed by CGH arrays were identical between diagnosis and relapse of T-ALL. We conclude that approximately one-third of late T-ALL “relapses” in fact represent second malignancies. We are currently performing further genomic analyses to identify common genetic events or common genomic features which might be related to predisposition for development of T-ALL.