In this issue of Blood, Müller and colleagues correlate clinical outcome on dasatinib with baseline BCR-ABL kinase domain mutation status, providing a framework for incorporation of BCR-ABL genotype in choosing the optimal second-line kinase inhibitor therapy in imatinib-resistant patients.1 This study should help improve outcomes by personalizing therapy based on the mutations detected.
Imatinib, the first approved BCR-ABL–selective tyrosine kinase inhibitor, was rapidly established as the preferred front-line therapy for newly diagnosed chronic phase chronic myeloid leukemia (CP-CML) on the basis of its ability to achieve a complete cytogenetic response (CCyR; defined as no detectable Philadelphia chromosome in at least 20 evaluable bone marrow metaphases) in the majority of cases. Unfortunately, it is estimated that imatinib fails to achieve CCyR within 18 months in approximately 25% of patients.2 These persons are at higher risk for the development of progressive disease, which has been most commonly associated with the development of drug-resistant BCR-ABL kinase domain mutations.3 Once the clinical importance of kinase domain mutations in BCR-ABL was recognized, 2 second-generation BCR-ABL kinase inhibitors (dasatinib and nilotinib) were clinically developed as they are active against nearly all imatinib-resistant mutations in vitro, with the notable exception of the T315I mutation, which is highly resistant to all 3 approved drugs in vitro.4,5
On the basis of preclinical studies demonstrating the T315I mutation to be the only substitution capable of conferring near-absolute resistance, it was hoped that this mutation would represent the principal clinical weakness of both dasatinib and nilotinib. However, there are clear limits to the predictive power of in vitro models. A true understanding of clinical resistance can be gleaned only from translational studies that interrogate appropriate clinical samples. Encouragingly, several studies have previously found the T315I mutation to represent the primary cause of loss of response to dasatinib, but other mutations at amino acid positions 299 (V299L) and 317 (F317L, F317I) can also be newly detected at the time of disease progression on dasatinib.6,7 Gratifyingly, initial preclinical studies demonstrated that of 15 tested mutations, F317L was second only to T315I in the degree of resistance conferred to dasatinib.4
To properly integrate preclinical observations into clinical practice, correlative studies of large patient cohorts are essential. To that end, Müller et al evaluated more than 800 patients enrolled in clinical studies with dasatinib following imatinib failure, and in 384 patients with detectable mutations at baseline, 63 different mutations were identified. Of mutations detected in at least 5 patients, 3 were associated with a particularly low probability of likelihood of CCyR achievement on dasatinib: T315I (0/21 cases), F317L (1/14 cases), and Q252H (1/6 cases). Similarly, a recent publication of imatinib-resistant CP-CML patients treated with nilotinib8 has revealed a handful of problematic mutations in addition to T315I (Y253H, E255K, E255V, F359C, F359V). Again, it is scientifically satisfying that these mutations are the most relatively resistant to nilotinib after the T315I mutation.5 Although neither report demonstrated a significant difference in outcome based on the presence or absence of any mutation at baseline, it is becoming clear that patients who harbor specific mutations may, in fact, be better served by treatment with a particular second-line agent.
A notable finding of the Müller et al study is the relative incidence of mutations that are potentially problematic to second-line therapies. Of the 384 patients with mutations, 42 patients had 1 of the 4 mutations documented to respond poorly to dasatinib, while 103 had mutations that do not respond well to nilotinib, a finding that is in general agreement with other studies that have assessed the relative frequencies of imatinib-resistant mutations in large cohorts of patients.9 This higher degree of cross-resistance between imatinib and nilotinib (relative to dasatinib) is not surprising given the degree of structural similarity between imatinib and nilotinib. However, it must be noted that no randomized controlled studies comparing dasatinib and nilotinib in patients with imatinib failure have been performed, and both agents achieve CCyR in a substantial proportion of CP-CML patients with imatinib resistance.
Nonetheless, the findings of Müller et al1 and Hughes et al8 provide clinicians with guidance regarding the preferential choice of a second-line agent. For imatinib-intolerant patients, there is no clinical evidence at this time to support either dasatinib or nilotinib, and weighing individual comorbidities with drug side effect profiles can be important. It is interesting that both Müller et al and Hughes et al identified BCR-ABL mutations in a small but significant proportion of imatinib-intolerant patients (8% and 10%, respectively), suggesting that some patients with imatinib-“intolerant” disease have “intolerant” disease have evolved subclinical imatinib resistance, and screening imatinib-intolerant patients, in addition to imatinib-resistant patients, for BCR-ABL kinase domain mutations may further guide clinicians in pursuit of the optimal second-line therapy.
Because CML represented the very first human malignancy to be treated with small-molecule tyrosine kinase inhibitors, it is hoped that the clinical experience with imatinib, dasatinib, and nilotinib will establish a paradigm that can be applied to this burgeoning class of molecules. With the recent work of Müller et al and Hughes et al, the era of personalized medicine for CML has arrived.
Conflict-of-interest disclosure: The author declares no competing financial interests. ■