Determining the Rise in BCR-ABL RNA That Optimally Predicts a Kinase Domain Mutation in Patients with CML on Imatinib.

Abstract

The most common mechanism of imatinib resistance in patients with CML is point mutations in the BCR-ABL kinase domain (KD) that impair imatinib binding to its target. As the second-line BCR-ABL inhibitors nilotinib and dasatinib show activity in most patients with KD mutations, early identification of mutations may prevent relapse. Current consensus recommendations are to perform mutation analysis if patients fail to achieve certain milestones of response, or experience loss of response. Since most patients are routinely monitored by BCR-ABL quantitative polymerase chain reaction (RQ-PCR), it is necessary to determine the optimal increase in BCR-ABL RNA that should trigger mutation testing, while minimizing both false-positive and false-negative test results. Recognizing the lack of a consensus, expert panels and the NCCN guidelines have provisionally recommended mutation screening in cases with a 10-fold or greater increase of BCR-ABL RNA. To address, in an unbiased fashion, the optimal transcript level rise associated with a mutation, and thus advise when mutation screening should be performed, 150 CML patients were serially monitored (median 46 months post-imatinib) by RQ-PCR and KD DNA sequencing. Thirty-five different mutations were identified in 53 patients, at a median of 28 months after imatinib initiation. During 22.5 months of follow-up after mutation screening, relapses occurred in 85% of the patients with a mutation (median 10. 4 months later) versus 46% of those with no mutation (median 32.4 months later) (hazard ratio = 3.2; P<0.0001). An unbiased receiver operating characteristic analysis identified a 2.6-fold increase in BCR-ABL RNA as the optimal cutoff for predicting a concomitant KD mutation, with a sensitivity of 77% for samples drawn at the same time as mutation discovery, and a sensitivity of 94% if including subsequent samples. At this 2.6-fold threshold, the negative predictive value was 97%, implying a small 3% risk of failing to detect a deleterious point mutation when the BCR-ABL RNA increase does not reach this optimized cutoff level. The 2.6-fold threshold approximated the analytical precision limit of our PCR assay. The table shows the diagnostic sensitivities, specificities, negative predictive values, and odds ratios for predicting a concomitant KD mutation at various rising RQ-PCR cutoff thresholds, including the optimal cutoff value of a 2.6-fold increase in BCR-ABL RNA. All cutoff values between a 2- and 4.5-fold increase were significantly predictive of a concomitant mutation, with 2.6-fold being the optimal cutoff level by several criteria (Youden index, negative predictive value, and odds ratio). In contrast, cutoff rises of 5- or 10-fold had poor diagnostic sensitivity (missing those 53–74% of mutation cases with lower levels of transcript rises), and were not significantly predictive of mutations. We conclude that a transcript level rise of >2.6-fold is the optimal cutoff to trigger reflex mutation screening in imatinib-treated CML. Furthermore, the currently recommended 10-fold threshold to trigger mutation screening may be overly stringent and is not universally applicable.RQ-PCR Increase (fold change)Sensitivity (95% CI) (%)Specificity (%)Negative Predictive Value (95% CI) (%)Odds Ratio (95% CI)P2.077 (62–88)4488 (80–94)2.6 (1.2–5.4)0.012.577 (62–88)4689 (81–94)2.8 (1.4–5.9)0.0052.677 (62–88)4789 (81–94)2.9 (1.4–6.0)0.0052.6* (include subsequent sample)94 (82–99)4797 (91–99)13 (3.9–43)<0.00013.074 (60–86)4888 (80–94)2.7 (1.3–5.5)0.0083.564 (49–77)5486 (78–91)2.1 (1.1–4.1)0.034.060 (44–74)6085 (78–91)2.2 (1.1–4.2)0.024.555 (40–70)6485 (78–90)2.2 (1.2–4.2)0.025.047 (32–62)6884 (77–89)1.9 (1.0–3.6)0.061026 (14–40)8382 (75–87)1.7 (0.8–3.7)0.2*Except for the one indicated row, the table includes only those patients with successful RQ-PCR and sequencing performed on the same sample. If samples after the initial genotype determination are included, at the optimal 2.6-fold cutoff, the sensitivity, NPV, and OR each increase.