Abstract
Background : Imatinib induces a complete cytogenetic response (CCR) in the majority of patients with chronic phase CML. CCR is durable in the majority of patients, but relapse occurs in a subset. To determine the potential of quantitative RT-PCR (qPCR) of BCR-ABL to predict cytogenetic relapse, we serially monitored residual disease in 90 CML patients with an imatinib-induced CCR.
Methods and patients : mRNA was prepared from total nucleated cells from blood or bone marrow, and cDNA was synthesized using random hexamer primers. Relative BCR-ABL expression was then measured by real-time fluorescent PCR normalized for G6PDH expression. This assay has a detection limit of 1 CML cell in 100,000 and an analytical precision of 6% (CV). At the start of imatinib therapy, 85% of patients were in chronic phase, at a median 9.5 months after diagnosis. Patients were treated with imatinib alone (64%) or in combination with interferon or cytarabine (32%). One patient each was treated with imatinib in combination with either the farnesyltransferase inhibitor tipifarnib, donor leukocytes (after allogeneic BMT), or an experimental heat shock protein (hsp70) vaccine. During the imatinib follow-up time of 28 months (median), disease monitoring occurred by cytogenetics and qPCR (median 6 samples per patient). The CCR was achieved after 9.7 months (median) of imatinib therapy.
Results : At the time of first achieving CCR, BCR-ABL RNA levels had decreased by a median of 1.8 logs below the median baseline level. During further follow-up, 26 patients (29%) experienced cytogenetic relapse (defined as any Ph-positive metaphase cell) at a median 6.0 months after CCR and a median 20 months after starting imatinib. There was no difference in the imatinib treatment time, the time to achieve CCR, or the post-CCR follow-up period between the patients with and without subsequent cytogenetic progression. qPCR data at the time of first CCR were available for 78 patients, including 25 of 26 with a subsequent cytogenetic relapse. The reduction of BCR-ABL RNA at the time of first achieving CCR was significantly less in those patients with a subsequent cytogenetic relapse (median 1.4 log) compared to those with a sustained CCR (median 2.0 log) (P=0.002). In the 64 patients with a sustained CCR, the molecular response progressively improved over time to reach a median reduction of 4.0 log at 15 months after CCR. Of the 29 patients achieving at least a 2 log reduction of BCR-ABL RNA at the time of first reaching CCR, only 3 (10%) had a subsequent cytogenetic relapse. In comparison, 22 of 49 patients (45%) with a less than 2 log reduction at the time of achieving CCR had a subsequent cytogenetic relapse (odds ratio = 7.1; 95% CI 1.9–26). At the time of first achieving CCR, a reduction in BCR-ABL RNA of less than 2 logs thus had a diagnostic sensitivity of 88% and a diagnostic specificity of 49% for predicting subsequent cytogenetic relapse.
Conclusions : We conclude that, in the majority of imatinib-treated CML patients reaching CCR, the level of BCR-ABL RNA at the time that the CCR is first achieved is a sensitive predictor of the durability of the CCR. The availability of a laboratory marker capable of stratifying the subsequent risk of disease progression (early in remission) will be useful in targeting additional (or alternative) therapies to those patients with the highest risk.
Author notes
Corresponding author
