AbstractAbstract 692 Background and Aims:In chronic myeloid leukemia (CML) and Philadelphia-positive acute lymphoblastic leukemia (ALL), tyrosine kinase inhibitor (TKI) therapy may select for drug-resistant Bcr-Abl mutants. Mutation status of resistant patients is usually investigated by Sanger sequencing (SS) of the Bcr-Abl kinase domain (KD). Novel ultra-deep sequencing (UDS) technologies allow to conjugate higher sensitivity with the unprecedented possibility to perform instant cloning of thousands of DNA molecules. We thus decided to take advantage of an UDS-based approach in order to: 1.resolve qualitatively and quantitatively the complexity of mutated populations surviving TKIs;2.investigate their clonal structure and evolution in relation to time and treatment. Methods:We retrospectively performed a longitudinal analysis of a total of 111 samples from 35 CML or Ph+ ALL patients who had received sequential treatment with multiple TKIs (two to four TKIs among imatinib, dasatinib, nilotinib, ponatinib) and had experienced sequential relapses accompanied by selection of TKI-resistant mutations. All samples had already been scored by SS; 74/111 (67%) were positive for one (n=33) or multiple (n=41) mutations. UDS of the Bcr-Abl KD was done using Roche 454 technology. UDS allowed to achieve a lower detection limit of at least 0.1% – as compared to 20% of SS. Results:Bcr-Abl KD mutation status was found to be more complex than SS had previously shown in 85/111 (77%) samples (representative examples are detailed in Table 1). In 33/74 (44%) samples known to harbour one or more mutations by SS, UDS revealed that up to four ‘minor' mutations with 1–20% abundance were present in addition to the ‘dominant' one(s). The type of mutations could easily be accounted for by TKI exposure history, since the majority were known to be poorly sensitive either to the current or to the previous TKI received. The higher degree of complexity was evident also when the clonal relationships of multiple mutations were reconstructed (Table 1). This revealed that identical mutations may be acquired in parallel by independent populations (e.g., one wild-type and one already harboring a mutation), via the same or different nucleotide changes leading to the same amino acid substitution (convergent evolution). In addition, longitudinal quantitative follow-up of mutated populations revealed that: 1.complexity generally increases with increasing lines of TKI therapy;2.with a few exceptions, double compound mutants have higher selective advantage over single mutants but also over triple;3.however, whether a compound mutant will ultimately take over depends on TKI, treatment duration, competition with other coexisting populations - the same compound mutants behaved differently in different patients receiving the same TKI. Conclusions:1.sequential changes in the selective pressure exerted by TKIs may result in a heterogeneous mosaic of subclones harbouring different mutations or mutation combinations;2.the evolution of each subclone is shaped not only by its inherent sensitivity to the specific TKI administered (‘absolute' fitness) but also by the competition with all other coexisting subclones (‘relative' fitness); (nonlinear) acquisition of additional mutations dictates further dynamics of shrinkage/expansion over time;3.information provided by SS may not always be sufficient to predict responsiveness to a TKI;4.sensitivity of a single or a compound mutant to a TKI in vivo is dictated by more complex factors than the mere in vitro IC50 value.Table 1SampleTKI (line)Mutations by SS (%)Mutations by UDS (%)Mutated populations by UDS (%)CML-04DAS (3rd, after IM and NIL)T315I (∼60) E255K (∼30)T315I (54.89) E255K (26.15) E255V (1.37)T315I (47.02) E255K (18.63) T315I+E255K (7.52) E255V (1.02) T315I+E255V (0.35)ALL-09PON (3rd, after IM and DAS)E255K (∼70) T315I (∼50)E255K (76.25) T315I (52.19) Q252H (cag>cac) (14.20) Q252H (cag>cat) (7.49) G250E (0.95)E255K+T315I (51.60) E255K+Q252H (cag>cac) (13.94) E255K+Q252H (cag>cat) (7.34) T315I (5.75) E255K (2.52) E255K+ G250E (0.70) Q252H (cag>cac) (0.25) G250E (0.25) Q252H (cag>cat) (0.15)CML-21DAS (2nd, after IM)Y253H (∼60) F317L (∼60)Y253H (54.90) F317L (54.40)Y253H+F317L (43.00) Y253H (11.90) F317L (11.40) Disclosures:Soverini:ARIAD: Consultancy; Bristol-Myers Squibb: Consultancy; Novartis: Consultancy. Castagnetti:Novartis: Honoraria; Bristol Myers Squibb: Honoraria. Luppi:CELGENE CORPORATION: Research Funding. Rosti:Bristol Myers Squibb: Consultancy, Honoraria; Novartis: Consultancy, Honoraria, Research Funding. Baccarani:ARIAD, Novartis, Bristol Myers-Squibb, and Pfizer: Consultancy, Honoraria, Speakers Bureau. Martinelli:NOVARTIS: Consultancy, Honoraria, Speakers Bureau; BMS: Consultancy, Honoraria, Speakers Bureau; PFIZER: Consultancy; ARIAD: Consultancy.