Bayesian Optimization Assisted Multistep Continuous-Flow Synthesis of Olutasidenib

Abstract Olutasidenib is a potent, selective, and orally administered small-molecule inhibitor of mutant isocitrate dehydrogenase 1. The traditional synthetic methods suffer from several drawbacks including prolonged reaction times, safety concerns, and low yields. Herein, we describe a four-step continuous-flow synthesis of olutasidenib via oxidation reaction, Katada reaction, methylation reaction, and nucleophilic substitution reaction. The coil reactor enables a shorter reaction time and reduces the risk of hydrogen peroxide decomposition due to its enhanced heat and mass transfer. Meanwhile, Bayesian optimization is used to accelerate the optimization of the continuous-flow. The level of N-methylation product of 2-pyridone was effectively increased to 93% by using methyl trifluoromethanesulfonate as a methylating agent in hexafluoroisopropanol. This continuous-flow process, featuring a total residence time of 86 min and a total yield of 44.1%, substantially reduces the reaction time and improves process safety while also achieving a lower Process Mass Intensity (PMI) than batch reactions.