Minimal residual disease (MRD) testing is vital to risk assignment in acute lymphoblastic leukaemia (ALL). Quantification of patient-specific rearrangements of immunoglobulin and T-cell receptor (Ig/TCR) genes is the most standardised method in current use. Recent studies demonstrated that deletion in the IKZF1 gene is common and is prognostic of poor outcome. IKZF1 targets could offer potential for MRD monitoring in BCR-ABLnegative (neg) ALL. The 5' and 3' break points of this deletion are highly conserved, making it possible to design a universal PCR assay to amplify the fusion genomic sequence created by the deletion.

With the initial aim of developing an MRD assay we studied the incidence of IKZF1 Δ 4-7 in 161 consecutive patients enrolled on the on-going adult ALL trial UKALL14, aged 25-59 years. We carried out PCR using breakpoint-specific primers designed as close as possible to the breakpoint. In order to ensure that the screening PCR enabled sensitive detection of the deletions, serial dilution of a positive control SUP-B15 cell line diluted into non-tumoral DNA was analysed. IKZF1 Δ 4-7 could be detected with a sensitivity of at least 10-4 (0.01%). PCR bands could be readily allocated to “strong” or “weak” on visual inspection; all PCR positive signals were confirmed as IKZF1 Δ 4-7 by Sanger sequencing of the PCR products and the breakpoints mapped. The frequency of the deletion in the total population studied was 80/161 (50%); 23/35 (66%) in the BCR-ABL positive (pos) subgroup and strikingly 57/126 (45%) in the BCR-ABLneg subgroup. The high rate of IKZF1 Δ 4-7 in the BCR-ABLneg patients was unexpected, being approximately double the reported incidence BCR-ABLneg adults.

Real time quantitative (RQ)-PCR analyses to investigate the suitability of IKZF1 quantitation as the basis for an MRD assay was performed on 26 BCR-ABLneg cases, selected only on the basis of availability of follow-up material. Assays were assessed using the same criteria applied by EuroMRD for Ig/TCR quantification. On that basis, “limited testing” of PCR amplifications at the 10-2 (1%) and 10-4(0.01%) dilution points was performed on all 26 samples. To our surprise, only 5 samples gave acceptable data to qualify for a quantitative MRD assay. Notably, all the 21 cases that failed “limited testing” had yielded a weak PCR signal on initial screening, suggesting the PCR assay have detected intragenic deletions restricted to minor subclones.

Although subclonal IKZF1 gene alterations are well described, the apparent high frequency of these events in our cohort (21/26 tested for RQ-PCR assay) was surprising. To further analyse this, we performed MRD-based genomic quantification of all 26 diagnostic specimens using SUP-B15 dilution series. The percentage of ALL cells bearing IKZF1 Δ 4-7 in the putative 21 subclonal cases was <0.01% compared to >90% in the putative 5 major clonal cases. In total, 82% of BCR-ABLneg cases gave a weak PCR band at initial screening, suggesting that most of the Δ 4-7 in BCR-ABLneg adult ALL are subclonal.

In the 5 cases where a quantifiable IKZF1 Δ 4-7-based MRD assay could be developed, MRD analysis was performed on 12 follow-up samples and sensitivity and quantitative range of at least 10-4 (0.01%) and 5×10-4 (0.05%) were obtained, respectively. These parameters compare favourably with the performance of standard Ig/TCR assays. Notably in one sample MRD could only be quantified by the IKZF1 deletion approach (level= 7.79E-05, ∼ 0.008%).

IKZF1 lesions are not leukaemia-initiating events, hence the stability of this alteration during the history of ALL is not clear. To address this, 14 diagnosis and relapse paired samples were analysed. Three different patterns emerged: in 5 cases the IKZF1 Δ 4-7 at diagnosis was preserved at relapse; in 3 cases the lesion was newly acquired at relapse; and in 3 cases the deletion was lost at relapse, notably all these 3 cases had showed a weak PCR signal, suggesting the deletion which was lost at relapse was subclonal at diagnosis. Theoretically this observation suggests that IKZF1deletion-based MRD monitoring carries a risk of missing a resurgent clone.

In conclusion, IKZF1 Δ 4-7 can provide highly sensitive MRD assays and could be a potential candidate to add to the repertoire of currently available MRD markers. However, we have shown limited applicability due to many IKZF1 deletions occurring in putative subclones. The stability of these deletions in major clones remains to be determined.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.