Michael Deininger, MD, PhD

Expert Perspectives
Michael Deininger, MD, PhD

Decisions in the Clinic: Treating Patients With Chronic Myeloid Leukemia


Michael Deininger, MD, PhD

Practice Community
Salt Lake City, UT
Practice Niche
Chronic myeloid leukemia
Hospital and Institute Affiliations
Huntsman Cancer Institute University of Utah


Do any patient or disease factors affect whether you assign newly diagnosed patients with chronic myeloid leukemia (CML) to receive a particular tyrosine kinase inhibitor (TKI)? Do high-risk cytogenetics affect your treatment decision?


The first question is disease phase. Patients with accelerated phase or blast phase at diagnosis are rare in the developed world, but they do exist. In these patients we use a second generation (2G) TKI, such as dasatinib, nilotinib, or bosutinib, as first-line therapy.

Patients with lymphoid blastic phase at presentation can be hard to distinguish from patients with Philadelphia chromosome (Ph)-positive acute lymphoblastic leukemia (ALL). So we will combine a TKI with ALL-like induction, unless the patient is not a candidate for intensive chemotherapy. People have argued that dasatinib is preferable due to its inhibition of SRC family kinases (which would also apply to bosutinib), but this is not clear-cut.

Patients presenting with myeloid blast phase will be treated with a 2G TKI, with the selection depending on comorbidities. We make an individual decision about adding chemotherapy. Irrespective of the phenotype of the blast phase, we have patients evaluated for possible bone marrow transplant (BMT).

Fortunately, the overwhelming majority of patients are diagnosed in the chronic phase. In our practice we recommend a 2G TKI to patients with intermediate or high Sokal risk. We regard the presence of cytogenetic abnormalities in addition to Ph as high risk and manage these patients like patients with high Sokal risk. The selection of 2G agent is then dependent on patient’s comorbidities and personal preferences.

It’s impossible to list all the conditions that will sway us toward or away from a particular TKI. We avoid dasatinib in patients with lung disease, nilotinib in patients with diabetes or coronary artery disease, and bosutinib in patients with liver disease. We consider imatinib in low Sokal risk or multiple prohibitive comorbid conditions.

A novel element in the decision process is treatment free remission (TFR) as a therapy goal. There are data showing that TFR rates are comparable in patients who achieve a deep molecular response, irrespective of which TKI got them there. Given the higher rates of deep molecular response on 2G TKIs, this implies that the chances of getting to TFR are better with 2G TKIs.

For young patients, and especially young women who plan to have a family, this is a very important consideration. In other patients it is more difficult and essentially a competing risk situation. For instance, is it worth accepting the higher risk of blood clots with nilotinib for the sake of a higher likelihood of TFR? These are questions that can only be answered in a thorough, open discussion between patient and physician.


Is cost typically a major issue for patients you encounter? How do you advise patients about managing the long-term costs of TKI therapy?


Sometimes it is, but at Huntsman Cancer Institute we are fortunate to have access to outstanding clinical pharmacology support. Even if insurance coverage is poor or absent we typically succeed in obtaining the TKI of clinical choice for our patients.

There have, however, been cases where we went with choice #2 to avoid a long delay. We have, for instance, started patients with high Sokal risk on imatinib to get them on treatment. If we are not happy with the response at 3 months we switch to a 2G TKI, then usually without problem.

TFR is a factor affecting discussions about long-term management of drug costs. I do not have all the answers here. The costs for generic imatinib need to come down, so that all patients have a medically acceptable, safe, and inexpensive long-term option. This amounts to basic health care, and in a civilized society everyone should have access to this, without the risk of financial ruin.


If patients develop resistance or intolerance to multiple TKIs, what other treatment strategies do you use?


Rule #1 is to max out every line of TKI therapy, which means not to switch easily and to manage adverse effects with appropriate supportive care. Above all, manage patient expectations from the very beginning. TKIs have side effects, but they are preferable to morbidity from CML or the consequences of BMT.

Intolerance and resistance can be hard to separate. As physicians we have the obligation to maximize tolerability so that the patients get the maximum therapeutic benefit from each TKI. Of course there are cases of real resistance to multiple TKIs, despite optimal compliance, and in these options are limited.

It is also important to define therapy goals. For instance, a partial cytogenetic response may be completely acceptable in an older patient with many comorbidities, although technically fulfilling the criteria for TKI failure. On the other hand, young patients who have failed ponatinib should be offered BMT, and this should happen as long as they are in chronic phase.

Alternatively we offer a clinical trial. Fortunately development of TKIs for CML continues, and asciminib (previously ABL001) is a promising agent that targets the BCR-ABL1 kinase in a different way compared to currently approved TKIs. With all this it is clear that failure to multiple TKIs is a prognostically ominous situation that requires careful consideration by a center experienced in the management of these patients.


What do you advise pregnant women or males/females of reproductive potential before initiating TKI therapy? Can pregnancy be successful while taking TKIs?


The law of the land is the women must avoid pregnancy while on a TKI, no exceptions. Data are relatively sparse and of course retrospective, but they clearly indicate an increased risk for fetal malformation, particularly in the first trimester. Therefore effective contraception is absolutely critical.

There are no prospective studies in men either, but all anecdotal data suggest that fathering a child while taking TKIs is safe. As such we do not recommend sperm banking for men. In women of childbearing age a complicated situation arises. We do not recommend preservation of eggs, considering the relatively low success rate and costs. The approach needs to be individualized. We tend to prescribe a 2G TKI to increase the chances of TFR — see my response to the first question — then we follow patients as usual, but plan to attempt TFR as soon as possible. This is typically after 3 years of therapy and 2 years in a deep molecular response (greater than 4-log reduction of BCR-ABL1 transcript measured on the international scale).

Of course, optimal compliance is required to get there, and as such managing side effects is paramount. Only 30% of woman will get to TFR, even with a 2G TKI, so for the majority this means CML management around a 9-months pregnancy. We always involve our colleagues form the high risk pregnancy team early on. The typical situation will be an acceptable response, such as a major molecular response. Here we advise patients to discontinue TKI, wait for approximately one month before attempting to get pregnant. This one month seems sensible, but it is not strictly based on data. We monitor patients with monthly polymerase chain reaction (PCR) for BCR-ABL1.

If a pregnancy occurs we sit tight until the third trimester, when it is safe to re-start patients on a TKI if needed. If we see a rapid climb of BCR-ABL1 we consider initiating therapy with pegylated interferon-alpha, which is safe for the pregnancy, but has considerable side effects. If there is a loss of hematologic control and concerns for placental leukostasis we perform leukapheresis.

The simple answer is that the approach defies planning and needs to be individualized, with many day-by-day decisions. We advise patients without an acceptable level of disease control (eg, complete hematologic response only) against a pregnancy. We are also fully transparent that a loss of disease control during a pregnancy may mean that the patient may not reach the same level of response once treatment is re-initiated.


What molecular markers determine whether you allow patients to discontinue treatment? How long, for example, do patients need to show major molecular response before discontinuation is feasible? How and how often do you monitor patients thereafter?


We consider patients with at least 3 years of TKI exposure and at least 2 years of a deep molecular response for a trial of TFR.

Patients with prior resistance to any TKI or prior progression to accelerated or blastic phase are not candidates for a trial of TFR. Major molecular response is not enough; it needs to be deep molecular response. We monitor monthly for 6 months after stopping, then every 2 months for another 6 months. If there is no evidence of recurrence we go to 3 months intervals. I should mention that there was a recent update of the NCCN [National Comprehensive Cancer Network] guidelines stipulating tighter monitoring, but we have not completely implemented this.