Abstract 2378: Responses of AML patients to tailored drug regimens: monitoring cancer subclones by ultra-deep resequencing

Abstract

Abstract As part of our individualized systems medicine (ISM) program, personalized treatment options are provided to clinicians based on in-depth genomic and molecular profiling as well as ex vivo drug sensitivity and resistance testing (DSRT) of leukemia patients (Pemovska et al. Cancer Discovery, 2013). In chemorefractory AML patients (n = 17), the ISM strategy has resulted in up to 35% response rate when individually selected targeted drugs have been applied in patient treatment. The responses achieved have, however, been transient and patients have typically relapsed quickly. Here, we aimed to understand the molecular basis of such treatment failures by quantitating the kinetics of individual cancer subclones before, during and after targeted treatments, as well as at the time of relapse and disease progression. Longitudinal serial samples from 13 AML patients were studied at multiple steps during leukemia progression and drug response. Clonal evolution of leukemic subclones was studied by both exome sequencing to get genome-wide overviews of disease progression, as well as by ultra-deep (>10,000x) amplicon resequencing with unique molecular identifiers to identify rare clones carrying specific cancer-relevant mutations. Nine of the 13 patients (69%) had multiple clones by exome sequencing and displayed branching evolution. In five patients who received treatment with targeted inhibitors we observed a significant differential therapeutic response of the individual AML subclones during therapy. In some cases, this could be directly attributed to the molecular mechanisms of drug response and resistance, such as the loss of NF1 in a subclone leading to cytarabine resistance or the loss of the FLT3-positive subclone in a patient responding to sunitinib treatment. Despite a prominent drug response at the level of the subclone carrying the driver mutations, in all these patients a new subclone emerged that led to progression of the disease. In three of the patients, the dominant clone appearing at relapse was already detected as a minor subclone in the diagnostic sample by amplicon resequencing. Amplicon sequencing enabled us to detect these minor subclones (down to 0,5% frequency) that were missed by exome sequencing. The results suggest that relapses in AML may arise because the drug-resistant subclone exists already before the onset of therapy. Overall, it is necessary to quantify tumor evolution and drug responses at the level of cancer subclones. Ultra-deep resequencing can be used to monitor drug responses at the subclone level, even at very low frequencies. This could facilitate early detection of small subclones with important prognostic implications, as well as the design of intelligent combinations of targeted drugs that could block such subclones. Citation Format: Poojitha N Ojamies, Mika Kontro, Henrik Edgren, Pekka Ellonen, Sonja Lagstrom, Henrikki Almusa, Timo Miettinen, Samuli Eldfors, David Tamborero, Krister Wennerberg, Caroline Heckman, Kimmo Porkka, Maija Wolf, Olli Kallioniemi. Responses of AML patients to tailored drug regimens: monitoring cancer subclones by ultra-deep resequencing. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2378.