Low Level BCR-ABL1 Gene Fusion Detected By RT-PCR in Newly Diagnosed Childhood Acute Lymphoblastic Leukemia Does Not Predict Philadelphia Chromosome Positive Acute Lymphoblastic Leukemia at Relapse: A 10-Year Single Institution Study

Background

The Philadelphia chromosome t(9;22), a reciprocal translocation between chromosomes 9 and 22, results in the gene fusion BCR-ABL1, and occurs in 2-3% of childhood acute lymphoblastic leukemia (ALL). It is detected using cytogenetic and molecular techniques: karyotype, fluorescence in-situ hybridization (FISH) for t(9;22) and reverse transcription polymerase chain reaction (RT-PCR) for BCR-ABL1. Detection has implications for treatment, with the addition of tyrosine kinase inhibitors to chemotherapy regimens improving outcome.

Low level BCR-ABL1 transcripts have been reported in blood of healthy individuals. We have observed this finding in bone marrow in newly diagnosed ALL in the absence of the t(9;22) by karyotype or FISH. The significance of low level positivity at diagnosis has not been determined in the setting of childhood Philadelphia chromosome negative (Ph-) ALL. Here we report, for the first time, the molecular evolutionary characteristics of children and adolescents with low level BCR-ABL1 positivity found at diagnosis to relapse.

Methods

We reviewed 327 patients aged 0-17 years diagnosed with ALL or Acute Leukemia of Ambiguous lineage (ALAL) at The Children’s Hospital at Westmead, Sydney, Australia from 1 January 2010 to 30 June 2020. Those positive for the BCR-ABL1 gene fusion by RT-PCR, and negative for t(9;22) by karyotype or FISH were included. Demographics, cytogenetics at diagnosis and relapse, and outcome data were extracted from the medical record.

Qualitative BCR-ABL1 analysis was performed using multiplex RT-PCR, followed by nested PCR, on RNA extracted from diagnostic bone marrow (sensitivity 5x10^-6). If positive, quantitation was performed using real-time RT-PCR with results expressed as the ratio of BCR-ABL1 over ABL1 (sensitivity 1x10^-5). Each PCR included positive and negative controls.

Results

Of 313 (96%) evaluable patients diagnosed with ALL or ALAL at our institution in the study period, 54 (17%) were positive by RT-PCR for BCR-ABL1 in diagnostic bone marrow. Seven patients were excluded as they had Ph+ ALL-specific treatment after the detection of t(9;22) by karyotype, FISH or other methods.

Forty-seven (15%) children with Ph- ALL had low level BCR-ABL1 detected by qualitative PCR. Demographic and cytogenetic characteristics for these patients are summarized in Table 1. All were diagnosed with ALL, the majority (77%) of precursor B-cell lineage including 2 with infant ALL. The e1a2 transcript was identified in 43 (91%) patients, with other transcript types as follows: e4a2 in 1 (2%), e13a2 in 1 (2%), and splicing variants in 2 (4%).

BCR-ABL1 quantitation was performed in 43 (91%) and was quantifiable only in 12 (28%) patients, with a median of 0.0008% (range 0.0003 – 0.095%).

Forty-five (96%) patients were treated with Berlin-Frankfurt-Munster ALL chemotherapy regimens. The two infant ALL patients were treated on the Interfant06 trial. One received a bone marrow transplant (BMT) in first remission then died after relapse; the other relapsed and died before BMT.

Seven (15%) of 47 relapsed, occurring at a median of 21 months (range 2 – 41 months) after diagnosis. Characteristics of these patients are presented in Table 2. Four patients were tested for BCR-ABL1 by RT-PCR in relapse marrow samples; all were negative. No patient with low level BCR-ABL1 positivity at initial diagnosis was diagnosed with Ph+ ALL at relapse.

There was no difference in 5-year relapse-free (80% vs 83%, P = .451) or overall survival (86% vs 91%, P = .368) between children with low level BCR-ABL1 positivity (n=47) and those without (n=259).

Conclusion

BCR-ABL1 low level positivity detected by RT-PCR in the bone marrow of children with newly diagnosed ALL is a relatively common finding, and did not adversely affect outcome for patients treated for Ph- ALL using a contemporary risk-adapted approach. Importantly, this finding did not influence the molecular evolutionary characteristics at the time of relapse in our patient group.