PF271 USING NEXT‐GENERATION SEQUENCING TO PREDICT RESPONSE TO SALVAGE CHEMOTHERAPY WITH HIGH DOSE CYTARABINE AND MITOXANTRONE IN RELAPSED/REFRACTORY ACUTE MYELOID LEUKEMIA PATIENTS

Background: A substantial segment of patients with acute myeloid leukemia (AML) will relapse following an initial favorable response to induction therapy or will prove to be primary refractory to standard induction therapy. Presently, there is a lack of consensus in the field regarding the optimal salvage chemotherapy regimen, thus leading to a wide variation of practice among clinicians. In our center, the current practice in this setting is to administer a regimen combining high dose‐cytarabine and mitoxantrone (HiDAC/MITO). However, predicting which patients are more likely respond to HiDAC/MITO specifically and to salvage chemotherapy in general is challenging, and prognostic tools with accurate predictive capacity are still lacking. Aims: Assess the clinical outcomes of relapsed/refractory (R/R) AML treated with HiDAC/MITO, and determine via next generation sequencing (NGS) whether patients responding favorably to this regimen harbor a unique molecular signature. Methods: We studied all adult patients with R/R AML who were treated with HiDAC/MITO in our institution between the years 2008–2017. Demographic, clinical, and laboratory data were collected using our institution's electronic medical records system. To determine whether patients responding to HiDAC/MITO had a unique molecular signature predictive of clinical response, we assessed 54 myeloid disease related genes using targeted next‐generation sequencing with the Illumina TruSight Myeloid Sequencing Panel (San Diego, California, USA). Results: The study cohort consisted of 172 evaluable patients with a median age of 54 years (range 18–77). A hundred and forty‐four (84%) patients had de‐novo AML, 24 (14%) had prior myelodysplastic syndrome, and 4 (2%) had an antecedent myeloproliferative neoplasm. Seventeen (10%) patients had favorable risk cytogenetics, 121 (72%) had intermediate risk cytogenetics, and 30 (18%) had poor‐risk cytogenetics. Fifty‐one (32%) patients were FLT3‐ITD mutated and 41 (29%) were NPM1 mutated. Ninety‐three (54%) patients were primary refractory to induction chemotherapy, 44 (26%) experienced relapse following consolidation chemotherapy, and 35 (20%) relapsed following an allogeneic stem cell transplantation. The response rate to salvage HiDAC/MITO was 58%; 11 patients (6%) died during salvage therapy. In multivariate analysis favorable risk cytogenetics [Odds ratio (OR) = 0.34, confidence interval (CI) 95%, 0.17–0.68; P  = 0.002], and de‐novo AML (OR = 0.4, CI 95%, 0.16–0.98; P  = 0.047) were found to be independent predicting factors associated with a favorable response to HiDAC/MITO. Neither the FLT3‐ITD and NPM1 mutational status nor the indication for salvage therapy (primary refractory versus relapsed disease or post‐transplant relapse) significantly impacted on the response to HiDAC/MITO salvage. Preliminary data from an NGS analysis comparing bone marrow samples from a subset of responding patients with non‐responding patients suggests that the responding patient segment was specifically enriched for mutations in DNMT3A , IDH2 , and IDH1.Summary/Conclusion: HiDAC/MITO is an effective salvage regimen for R/R AML, and prospective patients have a high chance of clinical benefit regardless of previous transplant, FLT3‐ITD / NPM1 status, and indication for therapy. Patients with adverse cytogenetics or secondary disease may not benefit as much from this regimen. NGS analysis has the potential to designate patients more likely to respond to HiDAC/MITO in this setting.