Mutated populations rise and fall in dominance over time in relation to therapeutic intervention. Graphical illustration of the kinetics of mutated population abundances in 4 representative cases. Arrows indicate the time points at which UDS was performed. Patient IDs are as in Table 3. (A) At the time of first relapse, 7 distinct imatinib-resistant mutated populations were detected. Dasatinib treatment cleared these mutants as quickly as in 2 months, but just as quickly a pan-resistant T315I mutant was found to have emerged. The patient achieved a transient hematologic response after 1 month but lost it shortly after detection of T315I. The question mark indicates that no T315I had been detected by UDS at the time of switchover (at a coverage of 4527 reads, suggesting that either it was present in <1/5000 transcripts, or that it was acquired some time later). (B) At the time of first relapse, a single imatinib-resistant Y253H mutant that accounted for almost 90% of BCR-ABL–positive cells was detected by SS and UDS. After 3 months of dasatinib therapy, the patient had achieved a complete cytogenetic response (no Ph⁺ metaphases detectable in the bone marrow by standard chromosome banding analysis) although residual disease remained detectable at the molecular level (as assessed by real-time quantitative [RT-Q]-PCR for BCR-ABL transcript). Neither the Y253H (known to be substantially sensitive to dasatinib) nor other mutants were detectable any longer by UDS. After 9 months on dasatinib, the patient was found to have lost the cytogenetic response. UDS showed the coexistence of 3 distinct compound mutants where Y253H was coupled with 3 well-known dasatinib-resistant mutations (a T315I and an F317L resulting from 2 different nucleotide substitutions). It might be hypothesized that the original Y253H-positive cells were never completely eliminated by dasatinib and persisted at very low levels (undetectable by UDS) until they happened to gain a selective advantage again, although de novo acquisition of mutations by previously unmutated cells cannot be ruled out. Interestingly, Y253H and T315I were already detectable by UDS 1 month before. (C) At the time of second relapse, after 6 months of dasatinib therapy, a dasatinib-resistant F317L mutation was detected. UDS portrayed a complex scenario with 3 distinct populations where the former imatinib-resistant Y253H and the newly acquired dasatinib-resistant F317L were present alone and in combination, although the Y253H+F317L compound mutant quantitatively dominated over the F317L- and Y253H-positive ones. It might be hypothesized that the same mutation was acquired in parallel by independent populations (ie, one unmutated and one already positive for the Y253H). (D) After 6 months of second-line dasatinib treatment, during which the patient achieved a complete cytogenetic response but no molecular response, the former imatinib-resistant H396R mutant plus 2 additional populations, 1 harboring a dasatinib-resistant F317L and 1 harboring H396R+F317L, were detected. Three months later, the F317L had become the dominant one, while H396R+F317L and H396R had declined. During dasatinib therapy, a pan-resistant T315I was also acquired because at the time of switchover to nilotinib, a T315I together with H396R+T315I and F317L+T315I compound mutants were already detectable by UDS (4.44%, 0.76%, and 0.22%, respectively). As quickly as in 2 months, nilotinib treatment selected the T315I mutant that expanded, achieving almost full dominance. IM, imatinib; DAS, dasatinib; NIL, nilotinib.