PS1212 UPDATED RESULTS FROM PHASE I TRIALS ASSESSING A NKG2D CAR T‐CELL APPROACH IN RELAPSED/REFRACTORY ACUTE MYELOID LEUKEMIA AND MYELODYSPLASTIC SYNDROME PATIENTS

Background: CYAD‐01 cells are chimeric antigen receptor (CAR) T cells based on the NKG2D receptor fused to the intracellular domain of CD3ζ, and bind ligands (MICA/B, ULBP1–6) expressed by a large variety of malignancies, including acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS). Aims: A comprehensive clinical program was developed with the objective to define the optimal CYAD‐01 treatment in relapsed/refractory (r/r) AML/MDS patients (pts). Methods: The Phase I THINK trial (NCT03018405) assesses the safety and clinical activity of multiple CYAD‐01 infusions without preconditioning chemotherapy in various malignancies, including r/r AML, MDS and multiple myeloma (MM). 3 dose levels (DL) were evaluated in the dose escalation segment (3x10 8 , 1x10 9 and 3x10 9  T‐cells/inf.). 1 st cycle of treatment consists of 3 CYAD‐01 infusions every 2 weeks and a 2 nd cycle of 3 CYAD‐01 infusions with a 2 week‐interval in case of no progressive disease (PD) after the 1 st cycle. Additional cohorts aim to provide a dense treatment schedule at the same DL (3x 1 CYAD‐01 every week + 3x 1 CYAD‐01 every 2 weeks). The Phase I/II DEPLETHINK study (NCT03466320) evaluates the safety and preliminary efficacy of 1 CYAD‐01 infusion administered after preconditioning with cyclophosphamide and fludarabine to pts with r/r AML/MDS. A 2 nd cycle of 3 CYAD‐01 without preconditioning is administered in case of no PD after 1 st treatment. 3 DL (1x10 8 , 3x10 8 and 1x10 9  cells/inf.) and 2 intervals between preconditioning and CYAD‐01 infusion are evaluated in the dose escalation segment. Data as of February 20, 2019 are presented. Results: THINK study: 16 pts were enrolled in the dose‐escalation segment of the hematological cohort, now completed. 4 pts were enrolled in cohorts with the dense schedule. In total, 7 pts experienced grade (G) 3/4 treatment‐related adverse events (AEs). Cytokine release syndrome (CRS) occurred in 10 pts with only 2 G3 CRS (at DL2) and 1 G4 CRS (at DL3), which resolved with early tocilizumab treatment. No treatment‐related neurotoxicity AEs have been observed. Out of 12 pts who received at least 3 CYAD‐01 infusions in the dose‐escalation segment, 2 pts with MM did not have evidence of clinical response. For pts with r/r AML/MDS, 4/10 pts had overall response (OR) at D29 with 1 complete remission (CR) with partial hematologic recovery (CR h ) for 16 months in a r/r AML pt at DL‐1, 2 CR with incomplete hematologic recovery (CR i ) for 1 month in AML pts at DL‐1 and DL‐3, and 1 marrow CR (mCR) for 1 month in an MDS pt at DL‐3. 2 AML pts at DL‐2 had stable disease (SD) with bone marrow (BM) blast percentage decrease upon 1 st treatment cycle and further reduction upon 2 nd treatment cycle. Two other AML pts in DL‐3 achieved SD for at least 2 months. 2 AML pts did not show evidence of clinical response. DEPLETHINK study: 6 pts were enrolled in the first 2 cohorts which evaluated 1x10 8  cells/inf. In total, there were 3 G3/4 treatment‐related AE observed in the same pt. Five CRS occurred in 4 pts: 3 G1/2 after the 1 st infusion and 2 CRS during 2 nd cycle (1 G4 CRS and 1 G1), which resolved with tocilizumab treatment. At the 1 st cohort, 2 pts reached a stable disease (SD) at D36, allowing the initiation of the 2 nd cycle of 3 CYAD‐01 infusions. Summary/Conclusion: Altogether, results obtained to date in both phase 1 trials demonstrate the safety of CYAD‐01 with or without preconditioning chemotherapy in patients with r/r hematological malignancies. A promising OR rate of 40% at D29 was seen in pts with r/r AML/MDS in the THINK study. The DEPLETHINK study is still accruing.