Key Points• UDS demonstrated that BCR-ABL KD mutations detectable with conventional methods may just be the tip of the iceberg.• The information provided by conventional Sanger sequencing may not always be sufficient to predict responsiveness to a given TKI.In chronic myeloid leukemia and Philadelphia chromosome-positive acute lymphoblastic leukemia, tyrosine kinase inhibitor (TKI) therapy may select for drug-resistant BCR-ABL mutants. We used an ultra-deep sequencing (UDS) approach to resolve qualitatively and quantitatively the complexity of mutated populations surviving TKIs and to investigate their clonal structure and evolution over time in relation to therapeutic intervention. To this purpose, we performed a longitudinal analysis of 106 samples from 33 patients who had received sequential treatment with multiple TKIs and had experienced sequential relapses accompanied by selection of 1 or more TKI-resistant mutations. We found that conventional Sanger sequencing had misclassified or underestimated BCR-ABL mutation status in 55% of the samples, where mutations with 1% to 15% abundance were detected. A complex clonal texture was uncovered by clonal analysis of samples harboring multiple mutations and up to 13 different mutated populations were identified. The landscape of these mutated populations was found to be highly dynamic. The high degree of complexity uncovered by UDS indicates that conventional Sanger sequencing might be an inadequate tool to assess BCR-ABL kinase domain mutation status, which currently represents an important component of the therapeutic decision algorithms. Further evaluation of the clinical usefulness of UDS-based approaches is warranted. (Blood. 2013;122(9):1634-1648

Funding information BolognAIL, AIRCThe most frequent BCR-ABL1 fusion transcripts in chronic myeloid leukemia (CML) are the e13a2 (b2a2) and the e14a2 (b3a2) ones. In the imatinib era few studies addressing the prognostic significance of the BCR-ABL1 transcript type in early chronic phase CML have been published. Overall, these studies suggest that in e14a2 patients the response to imatinib is faster and deeper. To evaluate if the BCR-ABL1 transcript type (e13a2 compared to e14a2) affect the response to imatinib and the clinical outcome in newly diagnosed adult CML patients, 559 patients enrolled in 3 prospective studies (NCT00514488, NCT00510926, observational study CML/023) were analyzed. A qualitative PCR was performed at baseline: 52% patients had a e14a2 transcript, 37% a e13a2 transcript, 11% co-expressed both transcripts and 1% had other rare transcripts. The median follow-up was 76 months (95% of the patients had at least a 5-year observation). The complete cytogenetic response rates were comparable in e14a2 and e13a2 patients. The median time to MR 3.0 (6 and 12 months) and MR 4.0 (41 and 61 months) was significantly shorter for e14a2 patients compared to e13a2 patients, with a higher cumulative probability of MR 3.0 (88% and 83%, P < .001) and MR 4.0 (67% and 52%, P 5 .001). The 7-year overall survival (90% and 83%, P 5 .017), progression-free survival (89% and 81%, P 5 .005) and failure-free survival (71% and

NCT00510926, NCT00514488, NCT00769327, NCT00481052.

© F e r r a t a S t o r t i F o u n d a t i o nten informed consent. The main exclusion criteria were a Performance Status ≥2, uncontrolled serious medical conditions, and prior treatment with imatinib. Molecular response (MR) was evaluated on peripheral blood every 3 months until a major MR (MMR; MR 3.0) was achieved and confirmed, then at least every 6 months. MMR was assessed and expressed according to the International Scale, as a BCR-ABL1 transcript level ≤0.1%, corresponding to a 3-log decrease from the International Scale standard, in samples with more than 10,000 ABL1 copies. MR 4.0 was defined as a BCR-ABL1 transcript level ≤0.01%, corresponding to a 4-log reduction, in samples with more than 10,000 ABL1 copies. 15Stable MR 4.0 was defined by at least five evaluations with MR 4.0, with no value >0.01%, during the preceding 2 years of treatment. Molecular tests were performed in one center (Bologna, Italy) until other GIMEMA centers were standardized for the International Scale. Mutational screening of the BCR-ABL1 kinase domain was done, using conventional Sanger sequencing as previously reported, 16 only in the case of resistance or progression. The baseline relative risk was calculated using Sokal, EURO, and EUTOS. [17][18][19] Blast phase was defined according to European LeukemiaNet recommendations. 20 Atherosclerotic adverse events were defined as peripheral arterial occlusive disease, acute coronary syndrome (acute myocardial infarction, instable angina), chronic ischemic heart disease (stable angina), and ischemic stroke. Cardiovascular risk factors [hypertension, diabetes, hypercholesterolemia, prior cardiovascular disease, obesity (body mass index ≥30)] were identified at baseline, prior to nilotinib treatment, on the basis of the medical history, and no specific suggestions for their management were provided in the protocol. Since this study was designed before the publication of data evidencing the cardiovascular toxicity of nilotinib, evaluation of serum lipids was not planned in the protocol (thus potentially leading to an underestimation of hypercholesterolemia), and no information on tobacco abuse was collected. Descriptive statistics are provided. The cumulative probabilities of events and survival were estimated using the Kaplan-Meier method. Results PatientsThe baseline characteristics of the patients are shown in Table 1. Few patients were at high risk: Sokal 14%, EURO 1%, and EUTOS 3%. The median age was 51 years, with 27% of the patients aged 65 years or more. Thirty-six per cent of patients had one or more cardiovascular risk factors: hypertension 19%, obesity 18%, diabetes mellitus 8%, hypercholesterolemia 4%, and prior cardiovascular disease 3%. Outcome and patients' dispositionNo patient was lost to follow-up, which ranged between 60 and 82 months (median 75 months). The 6-year overall survival rate was 96% (Figure 1). One patient progressed to blast phase on therapy, at 6 months, and died of leukemia. Another two patients died in remission, while on nilotinib: one with...