Background: Myelosuppression, and especially severe and prolonged neutropenia, can make patients (pts) with acute myeloid leukemia (AML) susceptible to fatal infections. Idasanutlin (idasa), a MDM2 antagonist promoting p53 tumor suppressor activity, showed tolerable safety and preliminary clinical activity alone and in combination with cytarabine (C) in pts with AML in a Phase I/Ib study (NCT01773408; NP28679). In the Phase III MIRROS study (NCT02545283; WO29519) comparing idasa + cytarabine (idasa-C) to placebo-C in pts with relapsed or refractory (r/r) AML, idasa-C did not improve overall survival (OS) in TP53-wild type pts (Konopleva, EHA 2020). The proportion of pts with composite complete remission (CR) at end of induction (EoI) was higher with idasa-C vs placebo-C; however, response at EoI did not translate into OS benefit. Longer and more pronounced neutropenia was observed with idasa-C. Here we investigated the contribution of idasa exposure variability to those findings, in particular, the relationship between exposure and toxicity and efficacy in pts with AML. Methods: MIRROS and NP28679 have been previously described (Montesinos, Future Oncol, 2020; Martinelli, EHA 2016). Plasma idasa concentrations were collected during induction and consolidation in 407 pts with AML (MIRROS: 285 with r/r AML; NP28679: 122 with AML), and a population pharmacokinetic (PK) analysis was completed. For each pt, predicted average concentrations values (Cav) were computed, as the ratio of cumulative area under the curve over the duration of the induction treatment period (5 days). Tertiles of Cav (low, medium, high) were used to reflect variability in exposure among pts. Logistic regression models assessed correlations between the probability of occurrence of adverse events (serious AEs [SAEs], febrile neutropenia), the probability of being responders at EoI and idasa exposure. These PK/pharmacodynamic (PD) analyses were restricted to pts treated with the idasa spray-dried powder formulation. In addition, the relationships between concentrations of the p53 target macrophage inhibitory cytokine 1 (MIC-1) and idasa exposure was characterized using an indirect PK/PD model developed in 245 pts with solid tumors (NCT01462175; NP27872), AML (NP28679) or polycythemia vera (NCT03287245; NP39761). Results: Idasa exposure was highly variable with lower exposure mainly associated with higher body weight and male sex. In addition, pts with vomiting during treatment had reduced exposure. However, a large part of the interpatient variability remains mostly unexplained. The exposure-response analysis, conducted in 316 pts with AML, showed a higher risk of SAEs with higher idasa exposure (Cav) during induction (p=0.03). There was also an association between idasa exposure and occurrence of febrile neutropenia (p=0.022; Figure 1). However, due to exposure variability, reducing the dose from 300 mg twice daily to 300 mg daily would result in a similar risk of febrile neutropenia at the median exposures (approximately 40%). No association between idasa exposure and complete remission rates at EoI was observed (N=295, p=0.622; Figure 2). Dose-exposure related increases in MIC-1 were observed at all idasa doses investigated, starting at the 100 mg total daily dose. Pts with AML in the lower tertile of exposure (Cav: 2503-5797 ng/mL) showed lower maximum MIC-1 release from baseline, contrary to pts from the highest tertile (7981-14651 ng/mL). MIC-1 release is present regardless of TP53 mutational status, suggesting a systemic rather than tumor derived induction. No difference was seen between responders and non-responders at EoI. Conclusions: Combined clinical pharmacology data in pts with AML receiving idasa revealed 1) Variability in idasa exposure does not correlate with clinical outcomes in pts treated with idasa-C. 2) p53 engagement (i.e., MIC-1 biomarker release) is seen at all dose levels investigated; however, MIC-1 release does not correlate with clinical response at EoI or TP53 mutational status. 3) The risk of SAEs is higher in pts with greater idasa exposure. 4) Data do not support lowering the dose of idasa to mitigate the risk of febrile neutropenia while sustaining clinical efficacy at idasa-C doses investigated. Disclosures Jamois: Roche: Current Employment, Current equity holder in publicly-traded company, Other: Roche: Support of parent study and funding of editorial support. Anders:Roche: Current Employment, Other: Support of parent study and funding of editorial support. Beckermann:Roche: Current Employment, Current equity holder in publicly-traded company, Other: Support of parent study and funding of editorial support; Novartis: Current equity holder in publicly-traded company. Genevray:Roche: Current Employment, Other: Support of parent study and funding of editorial support. Mundt:Swissmedic (starting September 2020): Current Employment; Roche: Current equity holder in publicly-traded company, Other: Support of parent study and funding of editorial support. Petry:Roche: Current Employment, Current equity holder in publicly-traded company, Other: Support of parent study and funding of editorial support. Yang:Certara: Current Employment; Roche: Other: Support of parent study and funding of editorial support. Kassir:Certara: Current Employment; Roche: Other: Support of parent study and funding of editorial support. Schmitt:F. Hoffmann-La Roche Ltd: Current Employment, Current equity holder in publicly-traded company.