A Half-Log Increase in Quantitative RT-PCR for BCR-ABL (RQ-PCR) as a Trigger for Kinase Domain Mutation Analysis.

RQ-PCR for BCR-ABL is standard of care for monitoring patients with chronic myeloid leukemia on Abl kinase inhibitor therapy, and mutations in the kinase domain (KD) of BCR-ABL are the most frequent mechanism of acquired drug resistance. From a practical point of view, it is important to establish an RQ-PCR trigger for mutation analysis that is sufficiently sensitive, while avoiding a high rate of false positive results. In one study, a single 2-fold (0.3 log) rise in BCR-ABL transcript levels predicted mutations in 61% (1). However, the results were not confirmed in a similar study(2). We therefore analyzed the correlation between a slightly more conservative 0.5 log rise in RQ-PCR levels and mutation detection in an unselected cohort of imatinib-treated CML patients. A database search identified 273 samples (from 150 patients) with a half-log or greater RQ-PCR rise. Mutation analysis (by direct DNA sequencing) has been completed on 54 samples. BCR-ABL levels were higher in 44 samples with successful amplification and sequencing compared to 10 without sufficiently good sequencing quality [median log-drop 0.31 (range −1.2 − 2.4) vs. 2.2 (range 0.17 – 3.2), p<0.0001]. KD mutations were detected in 11 samples (25%) and included L248V, G250E, Y253H, E282G, E292V, T315I (n=2), F359V+H396R, V379I, F382L, F482S). A similar mutation rate was detected by using a trigger of a 1.0-log increase of RQ-PCR levels [4 mutations out of 19 eligible samples (21%)], but with this higher trigger level, 7 patients with mutations would be excluded from detection. We then analyzed the risk of relapse according to mutation detection. The 44 samples with a >0.5 log RQ-PCR increase were from 34 different patients. Eleven of these 34 patients had cytogenetic or hematologic progression at the time of the mutation analysis and 7 never achieved even a partial cytogenetic response (<65%). Of 16 remaining eligible patients, 9 subsequently relapsed at a median of 5.8 (range 2.6–17) months after mutation detection. Freedom from progression was shorter in patients with a half-log RQ-PCR increase compared to patients without (34 months from imatinib start vs. median survival not reached, p = 0.019).

Conclusion: In our hands, a half-log increase of BCR-ABL levels is associated with mutation detection in 25% of samples, yields a significantly increased risk of subsequent relapse, and may be used as a practical trigger for kinase domain sequencing. Analysis of additional samples is ongoing to determine the ideal trigger level that optimizes specificity and sensitivity.