P337: IKZF1 DELETIONS IN B-ALL: FROM ITS GENETIC BASIS TO DIAGNOSTIC ENHANCEMENT

Background:IKZF1 deletions are associated with an increased risk of relapse in patients with B-cell precursor acute lymphoblastic leukemia (B-ALL), and their accurate detection has great clinical impact. Aims: We illustrate the recombination map of IKZF1 deletions, with a focus on non-recurrent deletions (neither Δ1-8, Δ2-3, Δ2-7, Δ2-8, Δ4-7, nor Δ4-8). In addition, we evaluate whether rare deletions are associated with a dismal prognosis for B-ALL patients, and provide a substantial basis for the improvement of diagnostic methods based on MLPA and multiplex PCR (M-PCR) to identify IKZF1 deletions. Methods: We included pediatric and adult patients with B-ALL of four international cohorts. IKZF1 deletions were screened using MLPA, and non-recurrent deletions were verified by NGS. After mapping the breakpoints, we identified breakpoint clusters (BC) to inspect genetic signatures associated with these DNA breaks. We used MEME for the agnostic motif search and FIMO for the identification of motifs in individual samples. We also annotated the presence of filler DNA (additional nucleotides) or microhomologies at the deletion junctions. The sequence data was used to update MLPA and M-PCR for the detection of IKZF1 deletions. Last, we compared the overall survival (OS) and cumulative incidence of relapse (CIR) of (1) non-recurrent IKZF1 deletions vs. (2) IKZF1 wild-type or (3) IKZF1 Δ4-7. The OS was compared by the log-rank test, while the CIR was analyzed by the Gray test. All data analyses were performed using R studio. Results: From 1,608 B-ALL samples, 17% had IKZF1 deletions. Non-recurrent deletions comprised 7% of them, and displayed a wide range of alterations: Δ1, Δ1-2, Δ1-3, Δ1-4, Δ1-5, Δ1-7, Δ2, Δ3, Δ4-6, Δ5, Δ5-7, Δ5-8 and Δ6-8. They included a total of 23 IKZF1 fusions. Most of them were reciprocal, leading to Δ1 and Δ1-3. After mapping the breakpoints, we defined 24 BCs. Four BCs (5’BC01, 5’BC02, 3’BC01, and 3’BC02) promoted rare rearrangements. The majority of ∆1 (9 out of 13) were associated with 3’BC01. This information allowed us to determine one commonly deleted region within IKZF1 exon 1, which is ideal for designing MLPA probes for detection of such deletions. In addition, it provided fundamental information to design an updated M-PCR to detect rare (∆1, ∆1-2, ∆1-3 and ∆2-3) and the most recurrent (∆2-7, ∆2-8, ∆4-7 and ∆4-8) deletions. Of note, our novel M-PCR extends the coverage of previous assays, which frequently missed the detection of ∆2-8 and ∆4-8. We also verified the occurrence of a heptamer-like sequence of RAG1 (E-value: 9.9 x 10-9) within BCs and an enrichment of GC nucleotides filling their deletion junctions (95% vs. 47% in deletions outside any BC). The three IKZF1 status groups had similar OS rates (P value = 0.60). On the other hand, we observed a higher 10-yrs CIR in patients with non-recurrent IKZF1 deletion compared to wild-type (14.7% vs. 50.1%; P value = 0.09). Summary/Conclusion: This study provides a wide spectrum of structural alterations that affect the IKZF1 gene, and point out methodological adjustments in MLPA and M-PCR to fine-tune the detection of IKZF1 deletions. Also, we summarize several levels of evidence that support the idea that RAG and TdT often mediate these alterations. Last, our results highlight that rare and recurrent deletions impact the prognosis of B-ALL patients. Therefore, their identification may contribute to appropriate risk stratification and therapeutic benefits for them. Keywords: Gene fusion, Ikaros, B cell acute lymphoblastic leukemia