Chronic myeloid leukemia (CML) is caused by increased tyrosine kinase activity resulting from BCR::ABL1 fusion. A chimeric mRNA transcripted from this fusion gene is translated into constitutively active tyrosine kinase and induces hyperproliferation of myeloid lineage cells, leading to leukemia. According to the breakpoints, fusion transcripts can be classified into several types with different molecular sizes. The breakpoint of ABL1 rarely shows variation and is located at the upstream region of exon 2 (a2). In contrast, the BCR gene has three breakpoint cluster regions (bcr): major, minor, and micro-bcr (M-, m-, and μ-bcr) [1]. Exons 13 and 14 in M-bcr are the breakpoints in the majority of patients with CML, resulting in fusion transcripts of e13a2 (b2a2) and e14a2 (b3a2), respectively [2]. The m-bcr is located upstream of the M-bcr and produces e1a2, e1a3, e6a2, and e6a3 transcripts. These transcripts are translated into lower molecular weight proteins than those of the major types. However, it is observed in 1.0% of patients with CML and is more closely associated with Philadelphia chromosome-positive B-cell acute lymphoblastic leukemia [3, 4]. The last and rare breakpoint, μ-bcr, is located downstream of M-bcr and gives rise to e19a2 (c3a2), e19a3 (c3a3), e20a2 (c4a2), and e20a3 (c4a3) transcripts, which are translated into higher molecular weight proteins than those of major-types [5].

Confirming the micro-BCR::ABL1 transcript, which originated from breakage in μ-bcr, is crucial for differential diagnosis because micro-type CML can express variable hematological presentations, including neutrophilic CML (CML-N), acute leukemia, and essential thrombocythemia [6, 7]. There are few longitudinal studies evaluating the incidence of patients with micro-type CML in Korea, and only several sporadic reports are available [8-10]. However, in the comprehensive analysis of BCR::ABL1 transcripts in Korean patients with CML, the micro-type incidence was 0.73% [11]. We report a Korean patient with CML expressing an e19a2 (c3a2) BCR::ABL1 fusion transcript. This work was approved by the Institutional Review Board of Severance Hospital, Seoul, Korea (IRB no. 4-2021-1566).

A 57-year-old man was referred to our center with leukocytosis (white blood cell: 35.6×10⁹/L), thrombocytosis (platelet: 1,637×10⁹/L), hepatomegaly, and weight loss. The initial differential complete blood count (CBC) performed at our center was as follows: neutrophil 72.2%, lymphocyte 12.0%, monocyte 2.5%, eosinophil 5.2%, and basophil 8.1%. A left shift in granulocytes (promyelocyte 2.0%, myelocyte 2.0%, metamyelocyte 3.0%, band 3.0%, segmented neutrophil 56.0%), a small number of blasts (2.0%), and increased eosinophils (7.0%) and basophils (7.0%) were noted in a peripheral blood smear (Fig. 1A). Bone marrow (BM) showed hypercellularity (>80.0%), and markedly increased small-sized megakaryocytes (Fig. 1B). The blasts in the BM were counted as 2.0% of all nucleated cells. There was no evidence of myelofibrosis or dysplastic features. The patient was suspected to have CML in the chronic phase and was started on hydroxyurea at 500 mg/day. The result of qualitative reverse transcription (RT)-PCR (HemaVision^® kit; DNA Technology, Aarhus, Denmark) analysis of BCR::ABL1 fusion showed negativity. This qualitative RT-PCR is a nested PCR that amplifies cDNA synthesized from RNA using two primer sets (first set: M6B PCR-I primers and M8F PCR-I primers,; second set: M6B PCR-II primers and M8F PCR-II primers). M6B primers cover six major- and four micro-type fusions, while M8F primers cover four minor-type fusions. JAK2, MPL, and CARL mutations were not detected. The patient was tentatively diagnosed with atypical CML, and the daily hydroxyurea dosage was increased to 1,000 mg. Approximately three weeks later, the patient’s platelet count surged to 2,123×10⁹/L. At this moment, the karyotype was reported as 46, XX,t(9;22)(q34;q11.2) [20]. The drug was promptly changed to a 2nd generation tyrosine kinase inhibitor (TKI), dasatinib at 100 mg/day. We reviewed the initial qualitative RT-PCR result and performed quantitative RT-PCR (qPCR using the Ipsogen BCR-ABL1 Mbcr IS-MMR kit; Qiagen, Hilden, Germany). When we carefully examined the results of gel electrophoresis, the internal control (IC) band looked thicker than those of other patients (Fig. 2A). After extending the electrophoresis run time up to 50 minutes, two distinct bands were revealed, a 911-bp IC band and a target band (Fig. 2B). During gel electrophoresis, we applied a molecular size ladder (Genepia, 100 bp plus #D1001-4, Korea) as a reference to estimate the size of the target band. The separated target band was located near the fourth band from the top (corresponding to a 1.0 kb size). As aforementioned, four micro-type fusions could be detected by performing nested RT-PCR using M6B primer sets: 1,147-bp-sized c4a2 type (fusion between BCR exon 20 and ABL1 exon 2), 1,012-bp-sized c3a2 type (fusion between BCR exon 19 and ABL1 exon 2), 973-bp-sized c4a3 type (fusion between BCR exon 20 and ABL1 exon 3), and 838-bp-sized c3a3 type (fusion between BCR exon 19 and ABL1 exon 3). We confirmed a fusion breakpoint of BCR exon 19 and ABL1 exon 2 using targeted RNA sequencing (Fig. 3), which refers to a 1,012-bp-sized target band. The initial quantitative RT-PCR analysis showed an international scale (IS) of 5.18% for the BCR::ABL1 transcript. After three months of TKI therapy, there was a 2.5-log reduction from baseline (IS 0.0149%). After six and nine months of treatment, IS values of 0.000225% and 0.000137% were measured, respectively; at 12, 15, and 18 months of treatment, negative results were observed for the BCR::ABL1 transcripts. In addition, the CBC complete blood count at 18 months of treatment was within the normal range, showing white blood cells of 8.03 ×10⁹/L, hemoglobin of 124 g/L, and a platelet count of 302×10⁹/L. Currently, the patient is steadily taking 100 mg/day of dasatinib.

RT-PCR is usually performed to identify the BCR::ABL1 fusion in CML due to its high sensitivity and relatively short turn-around time [12]. However, there is a caveat: the visualization of the IC band from the biotinidase housekeeping gene should be met. This serves as a verification of the RNA’s quality and the test’s efficiency. As μ-bcr is located downstream of M-bcr, a micro-type fusion transcript has a larger molecular size (838 bp to 1,147 bp) than other types [5]. Since the sizes of the IC band and the micro-type transcript band are similar, misinterpretation as a negative result can occur due to overlap. Therefore, it is essential to compare the thickness or pattern of the IC band with that of the negative control or that of other patients. If a micro-type fusion with a larger molecular size is suspected, the time of gel electrophoresis must be extended sufficiently, and correlation with conventional karyotyping and fluorescence in situ hybridization is highly recommended. In addition, targeted RNA sequencing that detects unknown partner genes and novel breakpoints using universal primers can be a powerful tool for obtaining accurate breakpoint information [13].

The detection of the micro-type variant is essential since it determines the initiation of TKI therapy. In our patient, provisionally diagnosed with atypical CML, improper treatment might have led to a rise in platelet count. The complete blood count dramatically normalized after the medication was changed to dasatinib, and our patient remains in remission. Nonetheless, there is a caveat in interpreting quantification results. A normalized copy number is obtained by dividing the BCR::ABL1 copy number by the ABL1 copy number, which is then converted to an IS value. Since an IS value has been established in patients with major-type transcripts, there are no standardized qPCR assays for monitoring molecular response to TKI treatment in patients with the micro-type transcript. For e19a2 (c3a2) type CML, qualitative RT-PCR or nested RT-PCR is generally recommended for monitoring minimal residual disease [14]. However, our patient was followed up by qPCR solely. Although our patient showed a low IS value at initial diagnosis, up to a 4.5-log reduction compared to baseline was noted using qPCR after TKI therapy.