Clinical responses to mIDH2 inhibition do not correlate with mIDH2 allele burden. (A) Dot plot of mIDH2 VAF (R140 mIDH2 in blue and R172 mIDH2 in red) in patient samples measured at screening in either peripheral blood or bone marrow by FoundationOne Heme panel. Measurements are separated by the best response achieved by patients, as defined in Figure 1. Numbers indicate the number of patient samples in the graph. (B) Waterfall plot indicating absolute change in mIDH2 VAF from screening to achievement of best response measured by Sysmex OncoBeam digital PCR. Responders are plotted in green and nonresponders in red. Patients achieving CR are outlined in black. The dotted line indicates the largest VAF decrease observed in a nonresponder. (C) Line graph of mIDH2 VAF over time (days of treatment) in 9 patients achieving molecular remission (undetectable mIDH2) for at least 1 time point during treatment. (D) Scatter plot of bone marrow mIDH2 VAF versus blast percentage measured by flow cytometry in 9 responsive patients in samples taken pretreatment (red) and at response (CR, CRi, CRp, or MLFS; green). Blue line indicates expected ratio (2:1) between blast percentage:mIDH2 VAF in clonal mIDH2 disease. Three data points (in green) are superimposed with values close to or at zero. BM, bone marrow; PB, peripheral blood; pre-Rx, pretreatment.