PS1058 NOVEL CIRCULATING TUMOR DNA (CTDNA) BASED DDPCR DROP‐OFF ASSAYS FOR IMPROVED DISEASE MONITORING IN ACUTE MYELOID LEUKEMIA (AML) A “LIQUID BIOPSY” FOR A “LIQUID TUMOR”

Background: While most AML patients achieve remission after induction therapy, relapse rates are high. Minimal residual disease (MRD) assessment has become an important parameter to determine relapse risk. Monitoring of MRD by repeated bone marrow (BM) aspiration has been shown to be more sensitive than analyses of peripheral blood (pB), yet it is invasive and associated with patient discomfort. Moreover, extramedullary disease (EMD) cannot be detected and intratumoral heterogeneity may be underrepresented in BM aspirates. Analysis of ctDNA in plasma may address these limitations. Strong correlation of mutation variant allele frequencies (VAF) in BM and ctDNA has already been reported in myelodysplastic syndrome, but low concentrations of ctDNA (<2–20.000 fragments/ml plasma) pose a challenge. Here, we report the use of highly sensitive digital droplet PCR (ddPCR) with newly designed drop‐off assays – allowing quantitative detection of diverse mutations at known hotspots – for disease monitoring in AML. Aims: Our aim was to evaluate the utility of ctDNA analysis as a novel sensitive tool for genetic diagnostics and disease monitoring in AML. Methods: Plasma samples were obtained from AML patients using Streck ctDNA BCT tubes, and ctDNA was isolated using the QIAampCNA Kit (Quiagen, Hilden, Germany) and quantified using the Agilent 2100 Bioanalyzer High Sensitivity DNA Kit (Santa Clara, CA, US). Mutations in IDH1 and NPM1 were quantitatively detected from ctDNA and matched BM and pB specimens using custom‐made ddPCR assays on a BioRad QX200 droplet digital PCR System (Biorad, Hercules, CA, US). All assays achieved a sensitivity of >99.9%. All patients provided written informed consent. The study was approved by the LMU medical faculty ethics committee. Results: Results from three illustrative AML patients are shown in the Figure.Patient 1 carried an IDH2 p.R172K mutation and received enasidenib. We compared serial ctDNA, pB and BM samples (Panel A) for disease monitoring. IDH2 mutation VAF in ctDNA closely correlated to BM VAF and was more sensitive than analysis of pB gDNA. These data suggest that ctDNA analyses may be useful for AML disease monitoring. Patient 2 had an NPM1 Type A mutation and received induction chemotherapy (Panel B). Despite rapid blast clearance in pB (by d4, not shown), ctDNA concentration in Plasma remained elevated and the NPM1 mutation remained detectable in ctDNA until day 15. These data imply that ctDNA in AML originates from BM and may be used to study response kinetics during induction therapy. Total ctDNA concentration, but not NPM1 VAF, increased again in parallel with WBC regeneration, implicating absolute ctDNA quantity as marker for high BM cell turnover. Patient 3 presented with isolated myelosarcoma, without morphological BM involvement. Biopsies were taken from the EMD site and BM. An IDH2 p.R140Q mutation was found in the EMD sample, but not in the BM. The IDH2 mutation was detectable in ctDNA at a VAF similar to the EMD biopsy specimen (Panel C). Hence, ctDNA allows mutational analyses of EMD, and potentially detection of extramedullary residual disease or relapse. Summary/Conclusion: Analysis of ctDNA by ddPCR assays allows sensitive, specific and less invasive genetic diagnosis and monitoring in AML. Our preliminary results indicate that mutation detection by ctDNA testing may be more sensitive than analyses of pB gDNA and may also capture EMD. We are currently investigating the prognostic value of measuring response kinetics during AML treatment. Recruitment of a larger patient cohort is ongoing, and further results will be presented at the congress.