Evaluation of BCR-ABL/ABL Ratio Increase That Corresponds to BCR-ABL Mutation In Chronic Myeloid Leukemia Patients Treated by Imatinib.

Currently there is no consensus in definition what level of BCR-ABL/ABL ratio increase predicts presence of kinase domain (KD) mutations. Several research groups use relatively low cut-off levels equal to 2.0- and 2.6-fold (S. Branford et al, Blood, 2004, R. Press et al Blood, 2009, respectively), that are close to the discrimination ability of real–time quantitative PCR (RQ-PCR) method. Alternatively, an NCCN guideline recommends beginning of mutation screening in case of 10-fold or greater elevation of BCR-ABL/ABL ratio.

To define a threshold level of BCR-ABL/ABL increase that predicts presence of BCR-ABL mutations.

Among 531 CML patients on imatinib (IM), both newly diagnosed and pre-treated with interferon-α, in 47 ones BCR-ABL mutation detection was performed. These were patients with suboptimal response or treatment failure according to the European LeukemiaNet criteria (M. Baccarani et al, 2009). Conventional cytogenetic analysis was performed every 6 months. Quantitative measurement of BCR-ABL/ABL transcripts ratio by RQ-PCR was done every 3–6 months. A major molecular response was defined as BCR-ABL/ABL transcripts level of 0.059% corresponded to 3 log reduction from the laboratory defined baseline level. Point mutations in the BCR-ABL KD were detected by reverse-transcriptase PCR and direct sequencing. Elevation of BCR-ABL/ABL was calculated by dividing of BCR-ABL/ABL value at the time point (TP) where mutation detection was performed to the BCR-ABL/ABL value at TP prior to mutation screening. Event-free survival (EFS) was defined as the time from IM beginning until any of the following events occurred: loss of complete hematological response, loss of major cytogenetic response, progression to AP/BC, death of any reason. Threshold level was defined by receiver operator characteristics (ROC) curve analysis. Positive and negative predictive values (PPV, NPV), sensitivity, specificity and overall correct prediction (OCP) were calculated.

10 different point mutations of BCR-ABL gene were detected, including 3 ones in P-loop, 2 in IM-binding site, 3 in A-loop, and 2 mutations outside the KD. None of patients had 2 or more mutations simultaneously. Patients were divided into two groups: with (n=18) and without (n=29) BCR-ABL mutations. Groups did not differ in age, sex distribution, type of BCR-ABL transcript, frequency of cumulative achievement of CHR, CCyR, MMR and level of BCR-ABL/ABL increase (table 1). Median time between BCR-ABL/ABL measurement was similar in both groups: 6 months (range 1–12 months) (p=0.227). ROC curve analysis determined that increasing of BCR-ABL/ABL level in 5.5-fold corresponds to 92.9% of NPV. Area under curve was 68% (95% CI 50–95%) (p=0.022). Sensitivity, PPV and OCP were relatively low (40.6%, 40.6%, 56.5%, respectively) while specificity was high (92.9%).

In our series 5.5-fold increase of BCR-ABL/ABL clearly predicts presence of BCR-ABL mutations and indicates the exact time for mutation detection performing in patients with suboptimal response and treatment failure. Nowadays, with availability of primary reference material for BCR-ABL quantification, approved by WHO (H. White at al, Hematologica, 2010) and successful harmonization of molecular monitoring of CML therapy (M. Mueller et al, Leukemia 2009) elevation level that corresponds with mutation presence could also be standardized. Application of international standardized threshold level would help to avoid unnecessary or late mutation tests.

Ivanets:Novartis Pharma: Employment.