In the approach to the management of newly diagnosed CP-CML, the major areas of clinical decision making include determining which of the available TKIs to start with, how to best treat the individual patient with the selected TKI, and whether and how to discontinue TKI therapy once a sustained deep response is achieved. As seen at ASH 2021, data relevant to all 3 of these important questions were presented.

Currently, we have multiple options for first-line therapy in adults, including the first-generation TKI imatinib and the second-generation agents nilotinib, dasatinib, and bosutinib, so the question becomes: Which one do you start with? Although the second-generation agents result in faster, deeper responses, there is no overall survival advantage. Prognostic scoring systems such as the Sokal Index and the Hasford scores were initially developed prior to the imatinib era but are good for predicting a cytogenic response. Patients with a low Sokal score are more likely to respond to imatinib, whereas higher-risk individuals tend to not respond as well. However, by and large, we are still left with several options from which to choose. When a patient has low-risk CML, decisions of whether to start treatment with imatinib are individualized based on the individual patient and their comorbidities and goals of care, as well as the side effects of imatinib therapy.

Nonetheless, it could be useful to have a simple way, with mathematical modeling, to estimate how well a patient might be expected to do on imatinib. At ASH 2021, Zhang and colleagues described a predictive scoring system for failure-free survival in patients with newly diagnosed CP-CML receiving initial imatinib (abstract 632). This model incorporates easily obtained variables based on the patient’s peripheral blood presentation and the European Treatment and Outcome Study Long-Term Survival score. It establishes 5 different patient groups with differing probabilities of 7-year failure-free survival with initial imatinib therapy. This is a very reasonable approach, but validation in other countries and in different populations is needed. Additionally, such predictive modeling for the second-generation TKIs would also be useful.

Once the TKI is selected, the next major question becomes: How do you best treat the individual patient on that treatment? Dose optimization is an interesting area in oncology. We perform phase 1 studies to identify a maximum-tolerated dose, which is the highest dose that can be safely administered to a patient. However, lower dosing is often associated with better tolerability and compliance. Several abstracts brought to light the question of TKI dose optimization, including the abstract by Sasaki and colleagues, who reported data from The University of Texas MD Anderson Cancer Center exploring treatment with low-dose (50 mg/day) or standard-dose (100 mg/day) dasatinib in newly diagnosed CML (abstract 631). The data are intriguing, as outcomes with the 50-mg dose and the 100-mg dose appear to be similar. Propensity score matching was used to identify 77 patients in each cohort without significant baseline differences. Limitations of the study include the small size and limited follow-up.

While this study does not replace a phase 3 study, this is a reasonable approach in lieu thereof, and it does give me confidence when determining which dose of dasatinib to use. I will still generally start patients at 100 mg per day, but, for those who may be older and/or when there are concerns about tolerability, I will go lower. Similarly, once a patient has been responding to TKI therapy, doses can be optimized to the individual patient’s side-effect profile if they are reporting toxicities. With serious adverse effects such as pleural effusions, patients need to stop taking their current drug and switch to another. But with subtle, low-grade, low-intensity, chronic toxicities that may lead to a lack of compliance and adherence, dose optimization can be very important in mitigating these toxicities and in improving compliance. This concept of less is more may apply to many of the TKIs. The less severe, more common chronic toxicities experienced with imatinib use may include myalgia, arthralgia, and diarrhea; with dasatinib use, it may be fatigue.

A different approach to dose optimization is taken with ponatinib in the TKI-resistant or -intolerant setting. Previously, participants in the phase 2 PACE trial were started at ponatinib 45 mg per day and stayed at that dose, and there was an accumulation of cardiovascular morbidity. This year, Jabbour et al reported that a response-based, dose-reduction strategy in the OPTIC trial mitigated much of the toxicity while improving the efficacy of ponatinib (abstract 2550). So, in this case, it was shown that if you start high and then go low, once you respond, you can achieve comparable or better efficacy while mitigating the risk of adverse events.

Regarding the deeper, sustained remissions that may allow for TKI discontinuation, developing predictive tools in this setting will also be crucial. Treatment-free remissions and treatment durations prior to discontinuation were explored at ASH 2021 (abstract 633). But there is not yet a model or scoring system that can take baseline factors and predict, with any precision, an individual patient’s likelihood of achieving treatment-free remission.