In reviewing CML drug development, Sweet and colleagues touch on the importance of dose optimization and lament that "TKI dose-finding trials… determine the MTD", and that "it is unlikely that most patients… require the MTD" (1). While we agree with the latter, dose optimization of the first TKI certainly did not stop at the MTD.Novartis' efforts to personalize imatinib dosing through Therapeutic Drug Monitoring (TDM) of plasma concentrations is unprecedented and never-repeated (2). The FDA's reaction disincentivized such efforts, and locked the industry into the one-dose-fits-all approach. Project Optimus merely suggests shifting a single dose, usually lower (3). We propose aiming for an optimal exposure range (therapeutic window), at whatever individual dose required. Lowering the dose will certainly reduce the number of patients above the therapeutic window; but other patients will move to below the therapeutic window, depriving them of benefit (Fig. 1), while the number of patients within the therapeutic window may not change. Reduced toxicity with equal activity may then be celebrated as a success. But optimal it is not. From the imatinib exposure response relationships (4), the CML field has now come full circle with ponatinib which has a 48% CV interindividual pharmacokinetic variability, corresponding to a roughly 5-fold range in exposures when everyone is given the same dose (5).The plasma exposure of a 15-mg ponatinib dose increase imparts odds ratios of 1.63 for MR1, and 2.05 for ≥MR1 response. The exposure increase of 15 mg imparts a HR of 2.05 for arterial occlusive events (AOE, causing its temporary market withdrawal) and 1.31 for grade ≥3 thrombocytopenia (5). Decreasing the dose for everyone reduces risk of toxicity for some, and reduces probability of response for others. The updated dosing for ponatinib (45 then 15 mg) merely reduces the time some patients are exposed to toxic concentrations and has a limited impact on OAEs. Dosing individual patients to a therapeutic window would have resulted in a distribution of doses beyond 15–45 mg, and might have been both more tolerable and more active.The authors point to idiosyncratic side effect profiles of TKIs with effects in some patients and none in others, and also stress the need for response biomarkers, suggesting early molecular response, T-cell subsets, NK-cell activity, or combinations. We suggest they revisit the TDM principles already established with their first revolutionary drug imatinib. Plasma exposure may be the best biomarker yet and therapeutic windows should be part of drug labeling.See the Response, p. 237S.J. Salamone reports other support from Saladax Biomecical, Inc. outside the submitted work. No disclosures were reported by the other author.