5.1 Imatinib

Imatinib mesylate was the first TKI to receive approval by the United States Food and Drug Administration (FDA) for the treatment of patients with CML-CP. It acts via competitive inhibition at the ATP-binding site of the BCR::ABL1 oncoprotein, which results in the inhibition of phosphorylation of proteins involved in cell signal transduction. It efficiently inhibits the BCR::ABL1 kinase and also blocks the platelet-derived growth factor receptor and the C-KIT tyrosine kinase.³⁶

The International Randomized Study of Interferon and STI571 (IRIS) in CML³⁷ randomized 1106 patients in CML-CP to receive imatinib 400 mg/day or IFN-a plus low-dose cytarabine. The outcomes with imatinib therapy were significantly better, notably the rates of CCyR (74% vs. 9%, p < .001), and freedom-from-progression to AP or BP at 12 months (99% vs. 93%, p < .001). Highlighting the challenge of using IFN-a was the high crossover rate to imatinib due to intolerance. With a median follow-up of almost 11 years, the estimated overall survival rate on imatinib therapy was 83.3% with a cumulative CCyR rate of 83% and a 10-year major molecular response (MMR or MR3; 3-log reduction of BCR::ABL1 transcripts, or BCR::ABL1 transcripts [IS] <0.1%) rate of 93%.¹⁸ Despite the high rate of crossover among patients assigned to receive IFN-a plus cytarabine (66%) and the short duration of therapy before crossover (median, 0.8 years), the 10-year survival rates favored the imatinib therapy arm (83.3% vs. 78.8%).¹⁸ In the German CML IV trial of 1551 patients treated in CP with imatinib-based regimens, with a median follow-up time of 10 years, the 10-year overall survival rate was 82%, the relative survival rate was 92%, and the cumulative incidence of CML-BP was 5.8%.⁶ Only 26.5% of patients switched to second-generation TKIs, 10% because of resistance, and the others for other reasons, including adverse events (AEs).⁶

5.2 Imatinib generics

Several groups have reported on the efficacy and safety of generic imatinib compared with Gleevec. The Polish Adult Leukemia Group (PALG) imatinib generics registry reported on 726 patients who started therapy with imatinib generics (multiple manufacturers) after the diagnosis of CML (n = 99) or who were switched to generic from Gleevec (n = 627) and were observed for at least 12 months.³⁸ Among patients treated in the frontline setting, the rates of 3-month PCyR, 6-month CCyR, and 12-month MMR were 66%, 53%, and 50%, respectively. These rates were similar to the expected rates achieved with Gleevec. The rates of resistance and intolerance were also similar (26% and 28%, respectively). Among patients who switched therapy, responses were maintained in the vast majority, with only 0.3% and 1% of patients losing their CCyR and MMR, respectively. In a study from India, among 174 patients treated with imatinib generics, response and survival rates were like those observed among 1193 patients treated with Gleevec. The safety profiles were similar as well.³⁹ Similar results were reported from our institution.⁴⁰

5.3 Dasatinib

Dasatinib is an oral, second-generation TKI that is 350 times more potent than imatinib in vitro.^41-43 It also inhibits the Src family of kinases, which may be important in blunting critical cell signaling pathways.²⁷

The DASISION trial was a phase III randomized study (1:1) comparing imatinib 400 mg once daily to dasatinib 100 mg once daily in 519 newly diagnosed patients with CML-CP.^28-30 The primary outcome was confirmed CCyR (cCCyR) at 12 months. Dasatinib therapy was associated with a higher cCCyR rate at 12 months (77% vs. 66%, p = .007). A 5-year follow-up showed that dasatinib induced more rapid and deeper responses (12-month CCyR, MMR, and MR4.5) at early time points compared with imatinib.⁴⁴ Transformations to CML-AP or CML-BP were fewer with dasatinib versus imatinib (4.6% vs. 7.3%). However, the estimated 5-year survival rate was similar (91% vs. 90%).⁴⁴

A second multicenter trial randomized patients with newly diagnosed CML-CP to dasatinib 100 mg once daily or imatinib 400 mg once daily.⁴⁵ Patients treated with dasatinib achieved a higher rate of CCyR (84% vs. 69%, p = .04). There was more toxicity experienced with dasatinib (Grades 3 and 4: AEs 58% with dasatinib and 35% with imatinib), mostly hematologic toxicity.⁴⁵

The superiority of dasatinib compared with imatinib was confirmed in a real-world retrospective evaluation in patients with newly diagnosed CML-CP.⁴⁶ Patients treated with dasatinib (n = 309) had higher rates of MMR (79% vs. 65%, p < .001) and deep molecular response (DMR; 44% vs. 25%, p < .001) compared with patients treated with imatinib (n = 304), with shorter median times to MMR (11.9 vs. 14.7 months, p < .001) and DMR (30.3 vs. 66.1 months, p < .001).⁴⁶

5.4 Lower dose dasatinib

The past experiences with cytotoxic cancer drugs developed and delivered these agents at a dose below the maximum tolerated dose; these were given intermittently and for short periods of time. The emerging BCR::ABL1 targeted therapies in CML (as well as later developed targeted therapies in cancer) were often administered daily and, because of their success in controlling CML and prolonging survival to a near-normal life expectancy, given indefinitely.^{17, 31, 47} However, the long-term follow-up uncovered AEs not previously reported.⁴⁸ This led to the concept of developing these targeted therapies in CML at an “optimal biologic dose”, defined hypothetically as a dose that maintains the same efficacy but reduces toxicities.¹⁷ This led to the study of dasatinib 50 mg daily as frontline therapy, which proved to be as effective and less toxic than 100 mg daily.^49-52 Among 83 patients with early CML-CP treated with dasatinib 50 mg daily,^49-51 the 3-month rate of early MR (BCR::ABL1 transcripts [IS] ≤10%) was 96%, and the 12-month MMR rate was 81%. The cumulative incidence of CCyR, MMR, and DMR (molecular response 4 [MR4] or MR4.5) at 5 years were 98%, 95%, and 82%, respectively. The rates of resistance (n = 4; 5%) and of toxicity (n = 4; 5%) were low. The 5-year overall survival rate was 96% and the event-free survival rate was 90%. No transformations to CML-AP/BP were observed. Grades 3 and 4 pleural effusions developed in 2% of patients.⁵¹

A propensity score analysis with 1:1 matching comparing frontline dasatinib 50 or 100 mg/day (77 patients identified in each cohort),⁴⁹ showed 3-year MMR rates of 92% and 84% for low-dose and standard-dose dasatinib, respectively (p = .23). Dasatinib 50 mg/day resulted in a higher cumulative incidence of MR4 (77% vs. 66%; p = .04) and MR4.5 (77% vs. 62%; p = .02) at 3 years. The 4-year failure-free, event-free, and overall survival rates were 89% versus 77% (p = .04), 95% versus 92% (p = .06), and 97% versus 96% (p = .78) with low-dose and standard-dose dasatinib, respectively. The rate of any grade pleural effusion was 5% with dasatinib 50 mg/day and 21% with 100 mg/day. Such a strategy of using more optimized OBDs of TKIs may have a significant impact on our future practice, mainly due to equivalent efficacy, a better safety profile, and a lower cost of care.

The DAVLEC phase II study from Japan evaluated dasatinib 20 mg daily as frontline CML-CP treatment in patients >70 years of age. Among 52 patients treated (median age 77.5 years; range, 73.5–83 years; median follow-up 12 months), 38 patients (73%) had BCR::ABL1 transcripts (IS) <10% at 3 months. The 12-month cumulative rates of MMR, MR4, and MR4.5 were 60%, 27%, and 14%, respectively. Three patients discontinued therapy due to treatment failure and one patient due to a drug-related adverse event (long QT syndrome). Four patients had pleural effusions (Grades 1 and 2).⁵³ The results from this study suggest that using dasatinib 20 mg daily is safe and feasible among older patients, often with medical comorbidities, allowing the achievement of molecular remissions with a lower incidence of AEs compared with higher doses.

5.5 Nilotinib

Nilotinib is a structural analog of imatinib. Its affinity for the ATP-binding site on BCR::ABL1 is 30–50 times more than imatinib in vitro.³²

The ENEST-nd was a large international randomized trial that compared two doses of nilotinib (300 mg or 400 mg twice daily [BID]) to imatinib 400 mg once daily.^{33, 54} The primary endpoint, MMR at 12 months, was significantly higher for both doses of nilotinib compared with imatinib (44% and 43% vs. 22%, p < .001). The cumulative incidence of CCyR by 24 months was 87% with nilotinib 300 mg BID, 85% with nilotinib 400 mg BID, and 77% with imatinib 400 mg daily (p < .001).^{33, 54}

With a median follow-up of 10 years,⁴⁸ the 10-year cumulative incidence of arterio-occlusive/vaso-occlusive events (AOEs/VOEs) was 24.8% with nilotinib 300 mg BID, 33.4% with nilotinib 400 mg BID, and 6.3% with imatinib. Consequently, nilotinib 400 mg BID should not be used as frontline therapy, although this dose schedule is the approved one in later-line therapies. The cumulative incidences of MMR by 10 years were 77.7% with nilotinib 300 mg BID and 62.5% with imatinib (p < .0001).⁴⁸ The 10-year cumulative rates of MR4.5 were 61% and 39.2%, respectively (p < .0001).⁴⁸ There was no significant difference in outcome among patients treated with nilotinib and imatinib.⁴⁸ The estimated 10-year progression-free survival rates were 86.2% with nilotinib 300 mg BID and 87.2% with imatinib.⁴⁸ The estimated 10-year survival rates were 87.6% and 88.3%, respectively.

In a second randomized trial with the same design that enrolled 267 Chinese patients, the 12-month MMR rate was 52% with nilotinib versus 28% with imatinib.⁵⁵ However, the rates of both CCyR (84% vs. 87%) and progression-free survival (95% each) were similar at 24 months. Overall, six patients in each arm progressed to AP/BP. In both arms, the estimated 2-year survival rate was 98%.

Other than the issue of AOEs with nilotinib, notable side effects were headache and skin rashes (common; 20%–30%—but mild to moderate; alleviated by dose reduction), self-limited elevation of indirect bilirubin (10%), elevations of blood sugar (10%–20%), and pancreatitis (1%–2%). Nilotinib can be de-escalated safely from 300–400 mg BID to 150–200 mg BID, or even 200 mg daily, if side effects occur or if there are safety concerns in patients who have responded optimally without compromising efficacy. This strategy is safe in the context of close monitoring.⁵⁶

5.6 Bosutinib

Bosutinib is a potent dual SRC/ABL kinase inhibitor.³⁴ The drug was first approved at 500 mg daily to treat CML-CP following resistance and/or intolerance to prior TKI therapy. It was recently approved at 400 mg daily as first-line treatment of CML-CP. In a multinational, phase III study, 536 patients with newly diagnosed CML-CP were randomized to bosutinib 400 mg once daily (n = 268) or imatinib (n = 268).³⁵ The 12-month MMR rate (primary endpoint) was higher with bosutinib (47% vs. 37%; p = .02), as was the CCyR rate by 12 months (77% vs. 66%; p = .0075). The favorable outcome with bosutinib was maintained at 2 and 5 years.³⁵ At 5 years, bosutinib resulted in higher cumulative rates of MMR (73.9% vs. 64.4%), MR4 (58.2% vs. 48.1%), and MR4.5 (47.4% vs. 36.6%) compared with imatinib. No additional transformation events (AP/BP) were recorded with this longer follow-up (six patients in the bosutinib arm and seven patients in the imatinib arm).⁵⁷ The 5-year overall survival rates were similar: 94.5% with bosutinib and 94.6% with imatinib. The notable side effects with bosutinib therapy are gastro-intestinal (diarrhea, but occasionally enterocolitis), hepatic (elevations of liver enzymes), and renal (increased creatinine). Such patient-associated organ dysfunctions should preclude bosutinib therapy in favor of other TKIs as frontline CML therapy.⁵⁷ We also recommend initiating bosutinib with a stepwise dose-adjusted approach, with dose escalation of 100–200 mg daily × 1–2 weeks, 300 mg daily × 2–4 weeks, then 400 mg daily, to avoid the early self-limited gastro-intestinal toxicities (diarrhea, nausea, and vomiting).^{16, 17, 34, 58} This strategy was evaluated in the salvage setting. In a phase II trial, bosutinib was shown to be safe and effective when given in a step-in dosing approach as second- or third-line therapy among CML patients intolerant and/or refractory to prior TKI therapy.⁵⁹

5.7 Investigating combinations of dasatinib with venetoclax or oral decitabine

While TKIs are effective in CML, they may not eliminate the dormant CML stem cells, which may result in disease relapse after treatment discontinuation.⁶⁰ The BCR::ABL1 tyrosine kinase regulates several BCL-2 family proteins that confer resistance to apoptosis in CML cells.⁶⁰ Targeting BCR::ABL1 and BCL-2 could have therapeutic benefits in quiescent CML CD34+ cells that are insensitive to TKI.⁶¹

The combination of a BCR::ABL1 TKI and venetoclax in clinical trials was aimed at increasing the rates of durable DMR and TFR. Unfortunately, the pre-clinical findings did not translate into positive clinical results.

Among 65 patients treated with the combination of dasatinib 50 mg daily with venetoclax added 3 months into therapy,⁶² the 12-month MMR, MR4, and MR4.5 rates were 86%, 53%, and 45%, respectively. After a median follow-up of 42 months, the 4-year event-free and overall survival rates were 96% and 100%, respectively. The outcome with the combination was comparable to the historical outcome with single-agent dasatinib, with a cumulative 12-month MMR rate of 79% with both strategies. Further follow-up is needed to evaluate the rates of durable DMR and TFR.

Decitabine was initially tested as monotherapy in 123 patients with CML (64 in BP, 51 in AP, and 8 in CP), with encouraging results.⁶³ The subsequent combination of decitabine and imatinib in patients with AP or BP was active and well tolerated. A phase I/II trial was conducted to determine the safety and efficacy of the combination of decitabine and dasatinib in 30 patients with CML in AP or BP. The respective rates of major hematologic response, major cytogenetic response (MCyR), and MMR were 48%, 44%, and 33%. The median overall survival was 13.8 months.⁶⁴ Based on these promising results, a novel combination of oral decitabine/cedazuridine (ASTX727) with low-dose dasatinib (50 mg/day) is under evaluation in newly diagnosed CML-CP (NCT05007873).

5.8 Other investigational combinations

In order to increase the proportion of patients who may be molecularly cured with long-term TKI therapy, several combination strategies are under evaluation. These focus on both inducing deeper MRs and targeting the CML quiescent stem cells. Several strategies are evaluating the addition of IFN-a, JAK2 inhibitors, anti-PD1 antibodies, and dendritic cell vaccines.^{5, 65-69}

The TIGER-trial (NCT01657604), a randomized phase III study, evaluated the efficacy and tolerability of nilotinib versus nilotinib + pegylated IFN-a (30–50 μg/week according to tolerability, initiated after ≥6 weeks nilotinib monotherapy) combination therapy for 2 years, followed by continuation of nilotinib in the standard arm versus IFN-a maintenance in the investigational arm. The study accrued 692 patients with newly diagnosed CML-CP.⁷⁰ Achievement of MMR, BCR::ABL1 ≤ 0.1% after >24 months of therapy was the trigger to start the maintenance phase; TFR was offered in patients with ≥12 months persistence of MR4 after >36 months of total therapy. The 24-month MMR and MR4.5 rates were 89% and 49% versus 93% and 64% with nilotinib versus nilotinib + pegylated IFN-a, respectively. In 273 (40%) eligible patients who discontinued therapy (nilotinib, n = 163; nilotinib + pegylated IFN-a, n = 110), the 2-year TFR rates were 53% and 59%, respectively. The 8-year progression-free and overall survival rates were 94% and 92%, and 95% and 94%, respectively. More AEs were encountered with the combination of nilotinib + pegylated IFN-a that impaired its tolerability.⁷⁰ One potential issue with the study design is using weaker later maintenance (IFN-a vs. nilotinib) in the investigational arm.

The addition of ruxolitinib 15 mg BID for 12 months to TKI in patients treated for at least 6 months with positive measurable l residual disease (BCR::ABL1 transcripts [IS] 0.0032%–1%) was evaluated in a randomized trial.⁷¹ Compared with TKI alone (n = 38), the combination of TKI plus ruxolitinib (n = 37) increased the 12-month MR4.5 rate (primary endpoint) from 3% to 15% (p = .09). The 12-month cumulative MR4 rates were 37% versus 63% (p = .048), respectively.

Current guidelines recommend any of the four TKIs, imatinib, dasatinib, bosutinib, or nilotinib, as good therapeutic options with a category 1 recommendation for the initial treatment of CML-CP. Second-generation TKIs produced higher rates of early optimal responses but had no impact on the long-term survival (probably because of available effective salvage therapies). They were also associated with higher rates of serious AEs (10-year cumulative rate of AOEs with nilotinib 300 mg BID 25%). Second-generation TKI may benefit patients with high-risk disease. Allo-HSCT or other chemotherapy agents are not recommended as frontline therapy for CML-CP, given the excellent outcomes and long-term survival achieved with the TKIs. Even considering the possible low cost of allo-HSCT as a one-time procedure in some geographies ($14 000–20 000 in India and Mexico), generic imatinib now costs less than $400–500 per year of therapy (total cost even with 40 years of life expectancy about $20 000), arguing for always delaying allo-HSCT to post-frontline CML therapy. Lower dose dasatinib (50 mg daily) showed similar efficacy and better safety compared with standard-dose dasatinib (100 mg daily). At MD Anderson, patients are currently offered clinical trials therapy with dasatinib 50 mg daily + oral decitabine or asciminib (discussed later). Because of the higher rates of durable DMR/complete molecular responses (CMR; undetectable BCR::ABL1 transcripts) with second/third-generation TKIs (which could lead to discontinuation of TKI therapy), considerations of second-generation TKIs in younger patients with CML (e.g., age < 50–60 years) may be considered if offered at a good treatment value (cost <$30 000/year), if inducing a TFR is the main goal of therapy. Otherwise, if the main goal of therapy is survival, any TKI would be appropriate. We favor generic TKIs, in particular generics of imatinib or dasatinib 50 mg (latter expected to be available in the US in 2025).

6.1 Aims of frontline CML-CP therapy

Selecting particular TKIs as frontline therapy depends on several considerations: (1) the aim of therapy, (2) the patient age and comorbidities, (3) the cost and affordability of the TKI under consideration, and (4) the CML risk profile. The two major aims of CML therapy are normalization of survival and possibly the achievement of a durable DMR (MR4 or MR4.5 lasting for 2–5+ years) and TFR status. For survival prolongation/normalization, at least 16 randomized trials have shown the equivalence of imatinib and second-generation TKIs, provided patients can afford the treatment, are treated optimally, are treatment compliant, and treatment is changed from imatinib to second-generation TKI promptly when there is evidence of CML resistance. If the aim of therapy is durable DMR/TFR, second-generation TKIs may achieve such results faster than imatinib. Thus, they may be considered in younger patients with CML-CP in whom TKI discontinuation and a TFR status are important.

6.2 Patient age, comorbidities, and TKI toxicity profile

While multiple TKIs are available to treat newly diagnosed CML-CP, each has a distinct toxicity profile to consider when deciding on therapy. Most TKIs are reasonably well tolerated with dose adjustments for toxicities, adequate monitoring, and supportive care.

Imatinib causes bothersome quality-of-life side effects including weight gain, fatigue, peripheral and periorbital edema, bone and muscle aches, nausea, and others. However, most are mild to moderate. Less than 5%–10% experience elevations in creatinine with long-term therapy. Rare neurologic toxicities were reported (worsening of Parkinsonism, dementia-like findings).

For patients at risk of developing pleural effusions (existing lung injuries), dasatinib should be avoided. This might be relevant for patients with a history of lung disease (e.g., chronic obstructive pulmonary disease), cardiac disease (e.g., congestive heart failure), or uncontrolled hypertension. Pulmonary arterial hypertension (PAH) is a rare yet important complication of dasatinib,⁷² and patients with preexisting PAH should be considered for alternative TKIs. Dasatinib also inhibits platelets function,⁷³ and patients taking concomitant anticoagulants may be at an increased risk of hemorrhagic complications.⁷⁴

Nilotinib has been associated with hyperglycemia; caution should be exercised in patients with uncontrolled diabetes when initiating therapy. Nilotinib should be avoided or prescribed with caution in patients with diabetes or a history of pancreatitis. During pre-clinical development, nilotinib was shown to potentially prolong the QT interval, and parameters were put in place to monitor for this complication after the drug was approved. Patients should take nilotinib in a fasting state to avoid excess drug exposure. Nilotinib has also been associated with AOEs, such as ischemic heart disease, ischemic cerebrovascular events, and peripheral artery occlusive disease (PAOD).⁴⁸ In the 10-year follow-up on the ENEST-nd trial,⁴⁸ 24.8% of patients experienced AOEs with nilotinib 300 mg BID. Nilotinib use should be limited or avoided in patients with risk factors such as diabetes or coronary, cerebrovascular, or peripheral artery disease. Avoiding nilotinib in patients with history of AOEs is warranted with the availability of other viable options. Bosutinib and imatinib are the safest TKIs in relation to AOEs.

Boustinib side effects are gastro-intestinal, hepatic, and renal. Among patients with such comorbidities, bosutinib should be avoided or used cautiously. Bosutinib should be avoided in patients with inflammatory bowel disease or renal dysfunction.⁷⁵ In order to avoid the early self-limited gastro-intestinal toxicities (diarrhea), it is recommended to start bosutinib at 100–200 mg daily × 1–2 week, then escalate to 200–300 mg daily × 2–8 weeks, then settle at 400 mg daily in the frontline therapy or at the optimal individual dose.^{16, 17, 58}

Finally, the patient age plays an important role in the treatment decision. Patients younger than 50–60 years are expected to live 30+ more years. Therefore, inducing a durable DMR may potentially lead to therapy discontinuation. Second-generation TKIs induce faster DMRs compared with imatinib. The issue of durable DMR and potential therapy discontinuation plays a less important role in older patients, in whom discontinuing therapy is less relevant.

With ongoing monitoring and evaluation, although rare, clinically relevant toxicities are emerging.¹⁷ Renal dysfunction and occasionally renal failure (creatinine elevations >2–3 mg/dL) are observed in 2%–3% of patients, more frequently with imatinib and bosutinib than other TKIs. This tends to reverse with discontinuation and/or dose reduction of the TKI. In rare cases, patients may develop TKI-related peripheral neuropathy or central neurotoxicity that are misdiagnosed as dementia or Alzheimer's disease. These may gradually reverse after TKI discontinuation.

While most side effects, even if severe, can be managed/reversed with dose reductions of a TKI, certain side effects can be prohibitive and require a change in TKI therapy (Table 2). Prohibitive toxicities include: (1) Recurrent pleural effusions—these are most commonly reported with dasatinib, less with imatinib and nilotinib; they respond to a short course of steroids. (2) Vasospastic or VOEs—these include cerebrovascular accidents, myocardial infarction, or unstable angina; they are more common with ponatinib and nilotinib and least common with imatinib and bosutinib. (3) PAH—noted with dasatinib (1%–2%), although it can rarely occur with other TKIs. PAH can reverse slowly after TKI discontinuation with a short course of steroids and sildenafil citrate. (4) Pancreatitis (2% with nilotinib, 2%–4% with ponatinib). (5) Neurologic problems (dementia-like, Parkinsonism)—these are rare and slowly reversible with TKI discontinuation. (6) Immune-mediated events (pneumonitis, myocarditis, pericarditis, hepatitis, and nephritis)—these are typically reversible with TKI discontinuation and a short course of high dose steroids (e.g., methylprednisolone 50 mg BID for 3–5 days). (6) Severe colitis—to watch for if bosutinib therapy is reversible with discontinuation.

When switching TKI therapy for prohibitive toxicity, the dose of the new TKI does not have to be the dose recommended for resistance. Generally, many of these patients have a good MR, and the dose of the new TKI can be lower, especially if the patient is already responding (≥MR2), for example, dasatinib 20–50 mg daily, bosutinib 100–300 mg daily, nilotinib 200 mg daily or 150 mg BID, and ponatinib 15–30 mg daily (Table 3).

TKI cross-intolerance may be more common than previously thought. When patients have intolerance/toxicities to one TKI, they may have more often intolerance to other TKIs. The intolerance/toxicity to the new TKI can be the same as with the previous TKI (e.g., a patient who develops a pleural effusion on dasatinib may be at risk of developing a pleural effusion on bosutinib), or a different one.

6.3 Cost and affordability of a TKI

The cost of cancer drugs has risen drastically over the past two decades.^76-78 All anticancer agents approved in the recent years were priced (average whole sale price [AWP]; what the patient pays on average) at over $200 000 annually.^{76, 79-82} Unlike the BCR::ABL1 TKIs in CML, most of these therapies do not provide a substantial improvement in survival or other objective measures to justify the cost. The BCR::ABL1 TKIs transformed CML from a fatal cancer to a chronic cancer with a near-normal life expectancy. The billionaire Mark Cuban created a generic company, CostPlus, that offers generic drugs directly to patients (thus avoiding the classical drug supply chains of intermediary suppliers) at cost +15%.⁸³ Therefore, a safe and effective generic imatinib is now available at $47 per month ($565 per year) in the US through CostPlus or generic companies that supply imatinib directly to patients. Considering the median age of 60 years in CML at diagnosis, the total cost of imatinib therapy for 40 years would be under $20 000. Since the long-term survival is comparable with imatinib versus second-generation TKIs, all patients with CML worldwide have access to affordable generic imatinib (at least 350-fold less expensive than dasatinib, nilotinib, or bosutinib). Raising awareness in the medical community of this “Mark Cuban effect” can spare patients the out-of-pocket expenses of second-generation TKIs, which can be as high as 20%–25% of the treatment costs.⁸³ Physicians will then have to assess the “treatment value” of second-generation TKIs (dasatinib, nilotinib, and bosutinib) in the frontline setting against generic imatinib in relation to benefits versus cost.⁸⁴ Second-generation TKIs (preferably generic) may be offered to patients with high-risk disease, while imatinib generic formulations are offered to patients with low- and intermediate-risk disease. Generic second-generation TKIs, offered at lower costs (<$10 000/year) may make them more cost effective as frontline TKIs therapy in younger patients based on different modeled scenarios considered.^84-86 Generic formulations of dasatinib are currently available outside the US. The formulations will be available in 2025 in the US, as will generic formulations of nilotinib, bosutinib, and ponatinib, hopefully by 2027. Currently, when survival is the endpoint, generic imatinib might be the best frontline TKI. When TFR is the endpoint, generic dasatinib as frontline provides a good treatment value.

The dosing of TKIs can affect the cost of therapy, depending on the TKI used. Using dasatinib at the lower dose of 50 mg daily is not only as effective and less toxic than 100 mg daily,⁴⁹ but also cuts the price by half, from an average AWP of $248 000 to $138 000 per year.⁸⁷ Unfortunately, ponatinib 45, 30, and 15 mg pills have the same AWP. However, in treating T315I-mutated CML, ponatinib AWP is $271 000 versus $1.8 million a year for asciminib 200 mg BID (the effective dose in T315I-mutated CML). However, neither ponatinib nor asciminib offer a good treatment value in this setting. For T315I-mutated CML, allo-HSCT (one-time procedure; cost range $20 000–500 000) may be the best curative option following a short course of ponatinib (3–6 months) to induce better disease control. While T315I-mutated CML-AP/BP is associated with a short survival, T315I-mutated CML-CP is associated with a median survival of 5+ years, allowing for a comfortable approach to ponatinib followed by allo-HSCT.⁸⁸

6.4 CML-CP risk status

For patients with CML-CP, several risk scores (Sokal et al.⁸⁹ or Hasford et al.⁹⁰) help predict outcomes. Patients with low- or intermediate-risk disease are expected to have optimal responses with imatinib, dasatinib, nilotinib, or bosutinib. However, second-generation TKI as frontline therapy may be more beneficial in patients with high-risk disease.^{44, 48, 57} Patients with higher-risk CML have a lower likelihood of achieving the early milestones of CCyR and MMR and may have higher risks of disease transformation to CML-AP or CML-BP.

Any of the TKIs currently approved for frontline CML therapy may be selected. These include imatinib, dasatinib, bosutinib, or nilotinib. While second-generation TKIs have demonstrated superiority over imatinib in relation to early surrogate markers, imatinib is still highly effective in most patients with CML. At MD Anderson, when choosing a BCR::ABL1 TKI, we consider issues such as aim of therapy (survival or TFR); cost of TKI and affordability to the patient; patient age, comorbidities, and TKIs AEs profile; and CML risk category. Kinase domain mutation profile plays no role in selecting an initial TKI but becomes relevant in the setting of CML resistance.

7.1 Monitoring

At baseline, all patients should undergo a bone marrow examination to establish the diagnosis, assess the percentage of marrow blasts and basophils, and perform a cytogenetic analysis to confirm the presence of the Philadelphia chromosome and to assess for ACAs, particularly i(17)(q10)-7/del7q, and 3q26.2 rearrangements.²⁴ The recommendation that patients have a follow-up bone marrow study at 3, 6, and 12 months after starting therapy, a staple of investigational trials, is not necessary in routine practice. An alternative method is with the use of FISH and PCR on peripheral blood.⁹¹ If a patient is responding optimally, and the BCR::ABL1 transcripts (IS) are <1%, marrow examinations can be omitted.⁹²

In the first year of therapy, monitoring peripheral blood by PCR every 3 months is reasonable. Once the patient is in a confirmed MMR (MMR for two to three times over a 6-month period), monitoring BCR::ABL1 transcripts every 6 months is adequate and safe, provided patient compliance is documented.⁹² For patients in CCyR (MR2), the achievement and maintenance of a DMR is of debatable significance for survival (unless the treatment aim is a durable DMR and TFR). Patients in CCyR/MR2 have similar survival regardless of whether or not they achieve MMR or DMR.^{93, 94}

7.2 Important points for determining treatment failure; significance of the “warning” and “failure” signals in the European LeukemiaNet recommendations/National Comprehensive Cancer Network guidelines

The European LeukemiaNet (ELN) recommendations and National Comprehensive Cancer Network (NCCN) guidelines have suggested and updated treatment milestones that define “warning” signals and “failure” (resistance) signals. These include “early MR” milestones (BCR::ABL1 transcripts [IS] <10%) at 3 and 6 months and later molecular milestones (MR2 at 12+ months; MMR at 36–48 months).⁹⁵ We believe these milestones may have confused the community practice into changing TKIs often and unnecessarily. With the maturing data, a simplification of the recommendations/guidelines may be indicated.⁹⁶

First, changing a TKI before or around the 3-month landmark is unwarranted unless there is clear-cut evidence of CML hematologic resistance (not just BCR::ABL1 transcripts [IS] >10%).^{91, 97-100} Second, while 6-month BCR::ABL1 transcripts (IS) >10% could indicate a need to change from imatinib to a second-generation TKI, such levels on a second-generation TKI should not lead to considering a third-generation TKI or allo-HSCT (unless a T315I mutation is documented).⁶ Third, based on the long-term follow-up results, the ELN “warning” signals should be eliminated, and the failure/resistance signals could be restated as “warning/caution” signals since they predict for an inferior outcome but not drastically so. Fourth, resistance, defined as BCR::ABL1 transcripts (IS) >1% at 12 months or later, is a clear indication of CML resistance and the need to change TKI therapy.⁹³ Fifth, changing TKIs in patients in MMR (or DMR) to achieve CMR (undetectable BCR::ABL1 transcripts) and TFR, or in MR2 or better to deepen the MR, are unwarranted and may cause more harm than benefit (new toxicities, increased cost, confusion). Sixth, in the particular case of older patients, persistent BCR::ABL1 transcripts (IS) 1%–10% at 2 years or later (who have exhausted all possible safe TKI permutations) may not warrant a consideration of allo-HSCT since these older patients may remain in chronic phase without CCyR for 10+ years on daily optimal TKI-based therapy (combined with hydroxyurea, cytarabine, hypomethylating agent, and omacetaxine).^{101, 102}

In several studies, the achievement of a CCyR (Ph-positive metaphases 0%; BCR::ABL1 transcripts [IS] ≤1%) at 12 months or later on TKI therapy was associated with a significant survival benefit compared with achievement of lesser degrees of response. Achievement of CCyR is the primary endpoint of TKI therapy. Achievement of BCR::ABL1 transcripts [IS] ≤0.1% (MMR) is associated with modest improvements in event-free survival rates, possible longer durations of CCyR, but not with a survival benefit. The achievement of durable DMR offers the possibility of treatment discontinuation and TFR. Lack of achievement of MMR or of DMR should not be interpreted as a need to change TKI therapy or to consider allo-HSCT. The NCCN and ELN updates of the criteria for resistance and suboptimal response have evolved over time, settling on more conservative criteria to change TKI therapy. The long-term follow-up studies have shown that patients who previously met early suboptimal or resistance criteria still had excellent long-term survival without changing TKI therapy. Therefore, in our opinion, these “warning” signals should be eliminated, and the failure/resistance signals could be restated as “warning/caution.” In contrast, resistance, defined as BCR::ABL1 transcripts (IS) >1% at 12 months or later, is a clear indication of CML resistance, and the need to change TKI therapy is warranted. A more conservative approach should be followed in older patients with persistent BCR::ABL1 transcripts (IS) 1%–10% at 2 years of imatinib therapy. These patients still had 10-year CML-specific survival rates similar to those with transcripts <1%.

8.1 Second-generation TKIs

Clinical studies of second-line TKIs are summarized in Table 5. Based on these studies, several noteworthy ideas have emerged. First, second-line treatment with dasatinib, bosutinib, or nilotinib can yield high rates of responses in patients who have inadequate response to imatinib, including high rates of MMR, and they are better than imatinib dose escalation.^106-110 Second, earlier switch to second-line TKI may be more effective than later switch. In the TIDEL-II study, patients who had suboptimal response to imatinib and were switched to nilotinib had a higher rate of DMR at 12 months than patients who had dose escalation of imatinib prior to being switched to nilotinib.¹¹¹ In a retrospective pooled analysis of three clinical studies of second-line dasatinib with CML with resistance or intolerance to imatinib, patients switched to dasatinib after the loss of MCyR (early intervention group) had higher rates of CCyR and MMR, as well as 24-month event-free, transformation-free, and overall survival than patients switched after the loss of both MCyR and complete hematologic response (CHR; late intervention group).¹¹² Although this analysis included studies with distinct study designs and various dosing schedules of dasatinib, the essential finding is that an earlier switch to dasatinib was associated with better outcomes.

Once a patient develops resistance to a second-generation TKI in second/later-line therapy (or even as frontline therapy), switching to another second-generation TKI is not indicated (yields low cytogenetic/MR rates) unless there is a guiding mutation. In such patients, changing to a third-generation TKI is better.

8.2 Third generation TKIs

Clinical studies of third-line TKIs are summarized in Table 6. Ponatinib is a third generation TKI, and the first TKI to exhibit activity against CML with T315I mutation.¹¹⁶ It is 500 times as potent as imatinib at inhibiting BCR::ABL1.¹¹⁷ The approval of ponatinib was based on the phase II PACE trial, in which 449 patients with heavily pretreated CML or Ph-positive acute lymphoblastic leukemia (ALL) were treated.¹¹⁸ Patients were considered for this trial if they had resistance to or intolerance of dasatinib or nilotinib, or if they had CML with T315I mutation. The dose of ponatinib was 45 mg once daily, and patients were stratified by disease phase and by the presence or absence of a T315I mutation. Among 267 patients who received ponatinib in CML-CP, 56% achieved a MCyR by 12 months, which included 45/64 (70%) patients with a T315I mutation. Patients responded more favorably if they had received fewer TKIs. After a median follow-up of 5 years, 60% of patients achieved MCyR (primary endpoint) at any time; 82% of those remained in MCyR at 5 years.¹¹⁴ Furthermore, 40% of patients achieved a MMR or better. The 5-year overall survival rate was 73%.¹¹⁴ AOEs occurred in 31% of patients (26% serious). The exposure-adjusted incidence of new AOEs (15.8 and 4.9 per 100 patient-years in years 1 and 5, respectively) did not increase over time. The most common all-grade treatment-emergent (TE) AEs occurring in ≥40% of CML-CP patients were rash (47%), abdominal pain (46%), thrombocytopenia (46%), headache (43%), dry skin (42%), and constipation (41%).¹¹⁴ Other notable toxicities include severe skin rashes (4%–7%), pancreatitis (7%), and severe hypertension (20%).

As of early 2014, ponatinib labeling included a revised warning regarding the risk of AOEs, heart failure, and hepatotoxicity.¹¹⁹ The AOEs were more frequent with increasing age and in patients with prior history of ischemia, myocardial infarction or prior vascular events, hypertension, diabetes, or hyperlipidemia, and with higher doses and with longer CML disease duration. Based on the results from the phase II OPTIC trial, an updated label was issued in December 2020, including an optimized, response-based dosing regimen of ponatinib, aiming to maximize efficacy while decreasing the risk of AEs.¹²⁰ In this study, 283 patients with CML-CP who failed prior TKIs (55% received ≥3 prior TKIs) or who had a T315I mutation (24%), were randomly assigned in a 1:1:1 ratio to receive a once daily dose of ponatinib 45 mg (45-mg cohort), 30 mg (30-mg cohort), or 15 mg (15-mg cohort).¹²¹ Upon achievement of BCR::ABL1 (IS) ≤1%, the dose of ponatinib was reduced to 15 mg daily. The 45 mg daily arm resulted in significantly higher response rates in T315I-mutated CML compared with the 30 and 15 mg doses (CCyR rate 60% vs. 25% vs. 10%, respectively), but less so in other CML subsets (CCyR rate 54% vs. 41% vs. 44%, respectively). The 4-year overall survival rates were similar with the three dose schedules at about 86%–88%.¹¹³ The incidence of AEs, serious AEs, and AOEs was significantly less than in previous studies. When comparing the dose modification dynamics of ponatinib from the PACE and OPTIC trials, the response-based dose-reduction strategy in OPTIC resulted in comparable or better efficacy outcomes, fewer dose reductions related to AEs, and fewer exposure-adjusted TE-AOEs, further demonstrating the benefit of the response-based dosing regimen used in OPTIC.¹²² We recommend using ponatinib 45 mg daily in T315I-mutated CML and 30 mg in non-T315I-mutated CML, with dose reductions to 15 mg daily once the BCR:ABL1 transcripts (IS) are ≤1%.

Although there were no head-to-head randomized comparisons in CML clinical trials with ponatinib to second and third generation TKIs, several studies demonstrated a higher potency of ponatinib compared with second-generation TKIs. In a systematic review of patients with CML-CP with resistance or intolerance to previous second-generation TKIs, treatment with a different second-generation TKI in the third-line setting was associated with a CCyR rate of 22%–26%. In contrast, treatment with ponatinib was associated with higher rates of CCyR of 60%.¹²³

A recent report assessing 354 patients with CP-CML (204 patients from the MD Anderson Cancer Center, 63 patients from the PACE trial, and 87 patients from the OPTIC trial) treated with third-line TKIs showed that ponatinib is the optimal treatment for patients who failed two prior TKIs.¹²⁴ In a multivariate analysis, ponatinib therapy in the third-line setting was the single independent factor associated with prolonged survival in CP-CML (hazard ratio [HR] = 0.45; p = .003). After propensity score matching, the 3-year overall survival rate was 87% with ponatinib versus 83% with second-generation TKIs (p = .03); the 3-year progression-free survival rate was 83% with ponatinib versus 59% with second-generation TKIs (p < .001).¹²⁴

The results of ponatinib therapy in the real-world setting are (surprisingly) superior to those reported in clinical trials. In an Italian study of 666 patients treated with ponatinib (median prior TKIs 3), among 515 patients treated in CML-CP, the MR2 rate was 78%, the MMR rate was 74%, and the MR4 rate was 43%. With a median follow-up time of 14 months, only 59/666 patients (9%) had died. The median time to discontinuation of ponatinib therapy in CML-CP was 47 months.¹²⁵

Based on these observations, ponatinib should be favored as the best TKI in patients with resistance to one second-generation TKI.¹²⁶ Table 7 summarizes different scenarios of TKI sequencing in patients with resistance.

Asciminib is an allosteric inhibitor that binds a myristoyl site of the BCR::ABL1 protein, locking BCR::ABL1 into an inactive conformation through a mechanism distinct from those for all other ABL1 kinase inhibitors.¹²⁷ In a phase I dose escalation study, asciminib was evaluated in 141 patients with CML in CP and nine patients with CML in AP with resistance and/or intolerance to ≥2 prior TKIs; 70% of the patients had received at least three TKIs. The recommended phase II dose among patients with CML-CP was 40 mg orally BID (or 80 mg daily).¹²⁸ In a 4-year follow-up among 115 patients with CML-CP without BCR::ABL1 T315I mutations and ≥2 prior TKIs, the most common grade ≥3 AEs included elevated pancreatic enzymes (22.6%), thrombocytopenia (13.9%), hypertension (13.0%), and neutropenia (12.2%). Clinical pancreatitis and AOEs occurred in 7.0% and 8.7%, respectively. At the last data cutoff, 61.3% of the patients with no CCyR at baseline achieved CCyR, 61.6% of those with no MMR at baseline achieved MMR, and 33.7% of those with no MR4 at baseline achieved MR4. Among patients with MMR at baseline and those who achieved MMR, 95% and 91% maintained their response with long-term follow-up, respectively.¹²⁹

The ASCEMBL phase III trial randomized 233 patients (2:1) with CML-CP post-failure of 2+ TKIs to receive asciminib 40 mg BID or bosutinib 500 mg/day. The primary endpoint was the incidence of MMR after 6 months of therapy, a new primary endpoint not confirmed to be associated with a survival benefit in CML salvage.^{17, 76} The 6-month MMR rate was 25.5% with asciminib versus 13.2% with bosutinib (p = .029). The MMR rate of 13.2% obtained with bosutinib is in contrast with the MMR rates of 50+% obtained in other bosutinib trials.¹³⁰ Based on these results, the FDA approved asciminib 40 mg BID for the treatment of CML-CP post-failure of 2+ TKIs without T315I or V299L mutations.¹³¹ The 12-month CCyR rates in patients without CCyR at baseline were 50.8% with asciminib and 33.7% with bosutinib. A longer follow-up of the ASCEMBL trial confirmed the sustained higher MMR rates with asciminib: 2-year MMR rate 37.6% with asciminib versus 15.8% with bosutinib (secondary endpoint of the study). However, this did not translate into improved progression-free survival (2-year rate 94% vs. 91%) or improved survival (2-year rate 97% vs. 99%).¹¹⁵ Fewer Grade ≥3 AEs were noted with asciminib compared to bosutinib (56.4% vs. 68.4%); the incidence of AOEs was 5.1% with asciminib. Among 39 patients who discontinued asciminib therapy because of lack of efficacy or disease progression, 10 (26%) acquired new mutations at the end of treatment, of which four (10%) affected the myristoyl pocket: A337T in three patients and P465S in one patient. No head-to-head studies comparing ponatinib and asciminib are available.

Asciminib was also approved for patients with T315I mutation based on a separate report, where 52 patients with CML-CP and T315I mutation who failed ≥2 prior TKIs received asciminib 200 mg BID (31 patients had prior ponatinib exposure, of whom 15 had ponatinib resistance).¹³² The estimated 2-year MMR rate was 57.8% in the ponatinib-naïve patients and 28.6% in ponatinib-treated patients. AOEs were reported in 5.8% of patients with the current follow-up.

Asciminib was evaluated in a Phase III trial in patients with newly diagnosed CML-CP. Patients were randomized to either asciminib or the investigator choice of imatinib or second-generation TKI. The primary endpoint was the rate of MMR at 12 months.¹³³ Asciminib (n = 201) induced a higher rate of 12-month MMR compared to investigator choice of TKI (n = 204); these rates were 67.7% and 49% (p < .001), respectively. Compared with imatinib (n = 102), these rates were 69.3% and 40.2% (p < .001), respectively. In contrast, when asciminib was compared with second-generation TKI, these rates were 66% and 57.8%. AOEs occurred in two (1%), zero (0%), and two (2%) of patients treated with asciminib, imatinib, and second-generation TKIs, respectively. Mutations occurring in or near the myristoyl pocket have been observed in eight of 201 patients (4%) treated with frontline asciminib (8 of 14 patients with lack of efficacy or disease progression, 57%), most commonly A337T, V506M, and L340Q, among others.¹³³ Even so, this may not automatically indicate that asciminib should be used as frontline therapy. For a new generation TKI to displace imatinib or second-generation TKIs as frontline therapy, it should meet (in our opinion) the following criteria (survival is unlikely to be demonstrably improved): (1) Achievement of higher rates of durable DMR (perhaps defined as rate of 2-year durable DMR). (2) Demonstration of long-term (at 8–10 years, not just 5 years) better safety profile than existing TKIs; this is important in light of the nilotinib experience, which reported no concerning safety signals at 5 years but showed a rate of AOEs/VOEs of 25%–35% at 10 years. (3) Good treatment value, that is, justifiable cost/year of therapy (perhaps <$20 000–30 000/year of therapy) for the benefits provided, considering that generic imatinib ($500/year; normalizes survival) and generic dasatinib ($5000–10 000/year; safe at 50 mg/day; high rates of durable DMR and TFR) are effective and meet the needs in CML therapy at low costs.

Asciminib was compared to ponatinib in a propensity score matching analysis among 426 CML patients. A total of 394 patients were included in the preliminary analysis: 254 with disease resistance and 126 with treatment intolerance; 89 patients had T315I mutation, and 89 patients had a history of cardiovascular disease. Treatment with ponatinib was associated with lower failure-free survival (FFS) compared to asciminib in the overall population; however, this difference was not significant after propensity matching. In the subgroup of patients without T315I mutation, those treated with asciminib had a higher FFS (p = .003), a higher rate of CCyR (p = .02), and a higher rate of MMR (p = .035) than those who received ponatinib. However, after propensity matching, only the CCyR rate was significantly higher in patients who received asciminib (p = .022).¹³⁴

Currently, there are no randomized head-to-head trials comparing ponatinib and asciminib in CML. The real-world results reported a CCyR rate of 78%, an MMR rate of 66%, and an MR4 rate of 43% with ponatinib given as third-line therapy.¹²⁵ These rates were 57%–70%, 22%–53%, and 16%–42% with asciminib, respectively.^135-138 Studies that compare ponatinib and asciminib in similar salvage settings (e.g., third-line therapy; T315I mutated or non-mutated CML) are needed.

8.3 Novel tyrosine kinase inhibitors

In November 2021, olverembatinib (formerly known as HQP1351),¹³⁹ a novel third-generation oral TKI, obtained regulatory approval in China for the treatment of TKI-resistant CML-CP and CML-AP with T135I mutations. In November 2023, it received approval in China for the treatment of CML-CP resistant or intolerant to first- and second-generation TKIs.¹³⁹ The approved dosage for olverembatinib is 40 mg every other day.

The safety and efficacy of olverembatinib were evaluated in (1) a single-arm phase I/II study in China in patients with TKI-resistant CP-CML and AP-CML;¹⁴⁰ (2) in a registrational randomized (2:1) phase II trial in China that compared olverembatinib to best available therapy (BAT);¹⁴¹ and (3) in a phase IB study outside China, evaluating the pharmacokinetics, efficacy, and safety of olverembatinib (three doses; 30, 40, and 50 mg every other day) in pretreated/refractory CP-CML and Ph-positive ALL, previously treated with ponatinib and/or asciminib.¹⁴²

In the phase I/II study, a total of 165 patients (>80.0% of whom had received ≥2 TKIs) were treated; of them 83% harbored T315I mutation (single mutation in 62%).¹⁴⁰ Among 127 patients with CML-CP, the 3-year cumulative rates of CCyR, MMR, MR4, and MR4.5 were 69%, 56%, 44% and 39%, respectively. Among 38 patients with CML-AP, the 3-year cumulative rates of CCyR, MMR, MR4, and MR4.5 were 47%, 45%, 39%, and 32%, respectively. Common treatment-related AEs included skin hyperpigmentation, hypertriglyceridemia, proteinuria, and severe thrombocytopenia.¹⁴⁰

In the registrational randomized phase II clinical trial conducted in China, olverembatinib was compared to BAT in 144 patients with CML-CP and resistance and/or intolerance to first- and second-generation TKIs.¹⁴¹ In this trial, 46% of the patients had ≥1 BCR::ABL1 mutation, including 27% with T315I mutation. Olverembatinib demonstrated superior CHR (85% vs. 34.8%), CCyR (36.4% vs. 16.2%), and MMR (27.3% vs. 8.1%) rates compared to BAT, along with a higher 2-year event-free survival rate (46.9% vs. 16.9%; p < .001).¹⁴¹

A recent study conducted in North America evaluated the efficacy, safety, and pharmacokinetics of olverembatinib in patients with relapsed/refractory CML-CP and Ph-positive ALL previously treated with ponatinib and/or asciminib.¹⁴² A total of 80 patients were treated; 51% had received ≥4 prior TKIs. Patients were randomized to receive olverembatinib 30, 40, or 50 mg orally every other day in 28-day cycles, with randomization stratified by T315I mutational status. The CCyR rate was 61% and the MMR rate was 42% in patients with CML-CP and were comparable between patients with or without T315I mutation. Among 26 patients with prior ponatinib exposure, 58% achieved CCyR and 37% achieved MMR. The rates of CCyR and MMR among patients with asciminib resistance were 50% (4 out of 8) and 33% (4 out of 12), respectively. In patients treated with ponatinib and asciminib, the CCyR rate was 43% and the MMR rate was 27%.¹⁴² The recommended phase III dose of olverembatinib was 30 or 40 mg every other day in patients with CML-CP without or with the T315I mutation, respectively. A randomized phase III trial in patients with at least 2 TKIs failure is ongoing comparing olverembatinib to bosutinib.

8.4 How to select a second- or third-line option

At the time of treatment failure, patients should undergo bone marrow examination to allow proper determination of the CML phase and documentation of any ACAs. All patients should have CML cells tested for BCR::ABL1 kinase domain mutations, as this will help guide the selection of the TKI.^139-143 When selecting between dasatinib, bosutinib, and nilotinib, in vitro and in vivo data have identified distinct mutations that exhibit decreased sensitivity to each of the agents.^143-146 Physician may favor dasatinib or bosutinib if the patient has the following mutations: Y253H, E255K/V, or F359C/V. Alternatively, nilotinib may be favored in the presence of the V299L and F317L mutations. For patients lacking these mutations, the choice should be based on preexisting conditions, toxicity profiles, and cost.

Although Sanger sequencing (SS) has been considered the gold standard for ABL1 kinase domain mutation screening, next-generation sequencing (NGS) has been increasingly evaluated in this setting. The NEXT-in-CML prospective study assessed the frequency and clinical relevance of low-level mutations and the feasibility, cost, and turnaround times of NGS-based BCR::ABL1 mutation screening in clinical practice.¹⁴⁷ A total of 236 patients with CML and treatment failure (n = 124) or warning (n = 112) were analyzed using SS and NGS. Fifty-one patients who were negative for mutations by SS had low-level mutations detectable by NGS. Furthermore, 29 of 60 patients who were positive for mutations by SS showed additional low-level mutations. Overall, 80 out of 236 patients (34%) had mutations undetectable by SS, of whom 42 (18% of the total) had low-level mutations of clinical relevance. These findings support the incorporation of NGS-based ABL1 kinase domain mutation screening results in the clinical decision algorithms.¹⁴⁷ ddPCR has been evaluated for the detection of ABL1 kinase domain mutation in CML patients with resistance to second-generation TKIs. Investigators used ddPCR to analyze 48 samples that were already studied by NGS and found that the ddPCR-based strategy was very accurate and specific, and more sensitive than NGS. It allowed the detection of ABL1 mutations with a 0.5% lower detection limit, irrespective of BCR::ABL1 levels. Test results were rapidly available, with a time from sample to results of 2 days.¹⁴⁸

In patients with resistance to a second-generation TKI without guiding mutations or with T315I mutation, a third generation TKI and consideration of allo-HSCT are indicated. The 5-year overall survival rate of patients in CML-CP treated with ponatinib and/or asciminib (90% treated with ponatinib and 10% with asciminib) was 77% compared to 50% with other therapies, including second-generation TKIs and chemotherapy.⁸⁸ A multivariate analysis showed that ponatinib and/or asciminib treatment (HR = 0.46; p = .03) was independently associated with better survival. Ponatinib is considered the preferred option based on the larger experience and longer-term follow-up with ponatinib compared with asciminib, the cost considerations (270 000 USD for ponatinib and 1 448 000 for asciminib 200 mg BID), and the safer dose-adjusted ponatinib schedules.

8.5 Allogeneic hematopoietic stem cell transplantation

The number of patients undergoing allo-HSCT for CML-CP has decreased significantly since the BCR::ABL1 TKIs became available, but may start to rise again as the prevalence of CML is increasing, and about 1%–2% of patients develop CML resistance to many TKIs every year and require allo-HSCT. Allo-HSCT has a more important role when patients evolve into AP/BP (see below).^{149, 150} Allo-HSCT remains an important therapeutic option for patients in CML-CP with resistance to a second-generation TKI (and without a guiding mutation) or who are potentially harboring the T315I mutation (after a trial of ponatinib or asciminib therapy).^{149, 150} Prior exposure to TKIs does not impact transplant outcome negatively; patients referred to transplant may have a better outcome if they undergo allo-HSCT in a better MR status (lower CML burden).^{151, 152}

Allo-HSCT cost versus TKIs cost and availability should be considered. Allo-HSCT, a curative one-time procedure costs $500 000+ in the US, but only $20 000 in some geographies. Allo-HSCT should not be offered as frontline therapy in view of the low-term benefits and low cost of available generic TKIs. In the salvage setting, the value of allo-HSCT versus second/third generation TKI can be discussed in relation to geography, cost of care, and availability of second-generation TKIs. For example, in some geographies, allo-HSCT as second-line therapy may cost $20 000 versus $228 000–290 000/year for second-third generation TKIs (excluding T315I mutation). In these circumstances, allo-HSCT could be offered as second-line therapy in nations with financial constraints.^{153, 154}

When should allo-HSCT not be considered post-TKIs failure/second-generation TKIs resistance? As experience accumulates, the dogma that “patients must achieve a CCyR or be considered for allo-HSCT” may not be as relevant in older patients (e.g., age >65–70 years). These patients may elect to forgo the option of a curative allo-HSCT in favor of a reasonably long and near-normal life with continuation of TKI therapy while not in CCyR.^{101, 102, 155} These patients may remain in a chronic phase on optimal TKI therapy alone or with the addition of other agents to maintain a hematologic response with or without a cytogenetic response.