Application of Real-Time PCR for the Detection of BCR-ABL1-like Group in Pediatric Acute Lymphoblastic Leukemia Patients

Background. In 2016 WHO classification of myeloid neoplasms and acute leukemias a new provisional entity 'B-lymphoblastic leukemia/lymphoma, BCR-ABL1-like' has been introduced. Two current strategies of BCR-ABL1-like ALL detection are based on gene expression profile revealed either by microarray ('BCR-ABL-like') or by NGS / patented TaqMan low-density array (TLDA) ('Ph-like'). Although both techniques and not widely applicable. Aims. To find out whether expression profile based on combined expression data of 5 genes assessed by real-time PCR can be used for the identification of BCR-ABL1-like pediatric ALL patients. Methods. The study was done on initial bone marrow samples of 147 primary pediatric BCP-ALL patients. Positive cohort included 10 BCR-ABL1-positive ALL patients, negative cohort consisted of 59 cases with known structural or numerical aberrations. Among them there were 21 patients with high hyperdiploidy, 1 low hypodploid, 1 near tetraploid and 1 iAMP21 cases, 23 patients with translocation t(12;21)(p13;q22), 5 cases with t(1;19)(q23;p13), 1 case with t(9;11)(p22;q23) and 6 Down syndrome (DS-ALL) patients. The rest 78 cases were called 'B-others' ALL and designated as training cohort. According to the previously published data (R. Harvey et al. ASH 2013 abs #826.) we picked 5 genes (IGJ, SPATS2L, MUC4, CRLF2, CA6) those expressions were estimated by real-time PCR. ΔCt method was applied in relation to ABL1 expression. In case of having expression results similar to BCR-ABL1-positive cases, patients with BCR-ABL1-negative ALL were attributed as BCR-ABL1-like ALL. In 113 patients we evaluated IKZF1 status by MLPA using P335 kit (MRC-Holland). Presence of ABL1, ABL2, CRLF2, IgH, JAK2, PDGFRb/CSFR1 gene rearrangements were done by FISH. Prognostic significance of BCR-ABL-like profile was estimated in 66 'B-others' patients uniformly treated according to the ALL-MB 2008 protocol. Informed consent was obtained in all cases. Results. Hierarchical cluster analysis and principal component analysis showed that 16 examined samples were clustered together with 9 BCR-ABL1-positive ones. Among them there were 3 DS patients, 1 iAMP21, 1 t(12;21) cases and 11 patients from 'B-other' group. IKZF1 deletions and high CRLF2 expression were more frequent in BCR-ABL1-like group in comparison to non-BCR-ABL1-like (67% vs 11%, and 40% vs 6%, correspondingly; p<0.001 in both cases). Among 9 BCR-ABL1-like cases which were assessed by FISH, JAK2 rearrangements were found in 2, CRLF2 in 4 (2 cases of CRLF2-IgH and 2 CRLF2-P2RY8). In 3 patients none of the tested gene rearrangements were revealed. We did not detect any tyrosine kinase fusion genes in 64 non-BCR-ABL-like 'B-other' patients (p<0.001). BCR-ABL1-like profile was associated with female gender (80% vs. 57% p=0.011), high initial WBC (≥30*10^9/L) (67% vs. 18%, p=0.001), M3 bone marrow status on day 15 of induction remission (27% vs. 4%, p=0.010). BCR-ABL1-like ALL patients more often were stratified to high-risk group (40% vs. 9%, p=0.001). EFS of BCR-ABL1-like patients was remarkably lower in comparison to non-BCR-ABL1-like 'B-other' patients enrolled into ALL-MB 2008 protocol (0.28±0.17 vs. 0.93±0.04, p<0.0001) while cumulative incidence of relapse was significantly higher (0.57±0.19 vs. 0.02±0.02, p<0.0001) with median of follow-up period 4.9 years. The worst outcome was noted in case of combination of BCR-ABL1-like profile and IKZF1 deletions (EFS 0.00), while patients with isolated BCR-ABL1-like profile doing much better (EFS 0.67±0.22). Poor EOI flow-MRD response (>0.1%) together with BCR-ABL1-like profile identified a group of patients with dismal outcome (EFS 0.00 vs 0.88±0.11, p<0.001). Overall accuracy of our BCR-ABL1-like profile for relapse prediction in 'B-other' group on ALL-MB 2008 protocol was 0.939 (95% CI 0.881-0.997), for risk of unfavorable event 0.909 (95% CI 0.840-0.978). Conclusion. Thus, we showed that described real-time PCR technology based on expression data of 5 genes allowed to detect the BCR-ABL1-like ALL patients with similar clinical characteristics, genetic parameters and treatment outcome to ones revealed by microarray, NGS or TLDA techniques. We believe that detection of BCR-ABL1-like profile by PCR can be used as fast and easy screening method in 'B-other' ALL for the further identification of tyrosine kinase fusion genes or other targetable lesions by NGS or FISH within this group only.