The impressive complete cytogenetic responses (CCyR) and the consequent long-term overall survival (OS) together with a good tolerability have rapidly attributed to the first BCR-ABL tyrosine kinase inhibitor, imatinib at a dosage of 400 mg per day, the role of standard treatment, and the most widely frontline therapy for chronic phase CML patients [17, 18]. At the 8-year follow-up of the IRIS study, as confirmed by other studies and by independent retrospective analyses outside clinical trials, the cumulative CCyR rate was of 83 % with estimated overall survival (OS) rate of 85 %, which is far better from what was observed before the introduction of this drug [19–22]. These results are also associated with the substantial decrease in the number of the progressions to accelerated phase or blast crisis imatinib-treated patients. In all reports, the major cause of death for CML patients is indeed the progression to the accelerated or blast phase, being still incurable in most cases even with tyrosine kinase inhibitors (TKI) [23]. During imatinib therapy, the progression occurrence falls from an expected rate of approximately 15 % per year to a rate of 2–3 % per year. Furthermore, after the first 2–3 years of treatment, the progression rates drop dramatically [19]. The reduction of progression depends mostly on the astonishing leukemic mass reduction which is observed in most of the imatinib-treated patient. Notwithstanding, the inhibition of the BCR-ABL tyrosine kinase (TK) activity affects the ability of BCR-ABL to promote the genomic instability of the leukemic cells which is a cause of disease progression [24].

Based on the new ELN and NCCN criteria of TKI therapy response evaluation, one-third of newly diagnosed CML patients do not show an optimal response to the standard therapy with imatinib 400 mg, and they are therefore facing a statistically significant higher risk of an inferior outcome in terms of EFS, PFS, and also OS (approximately 80 % at 5 years with respect to >95 % of those below 10 % BCR-ABL^IS at 3 months) [10, 14–16]. Actually we should consider that most of these patients (approximately 80 %) display only a delayed response to imatinib and that a switch to a second-generation TKI allows to achieve an optimal response in approximately 40–50 % of the cases [25, 26]. However, it should also be considered that approximately 15–20 % of them in a short time will progress to a more advanced phase of the disease and will die [10, 14–16]. In any case, several reports have shown that after 8 years from diagnosis, only approximately 55–60 % of the imatinib-treated patients are still on treatment with this drug [19, 20]. The reasons for imatinib discontinuation depend on failure, progression, and death; furthermore in about 10–12 % patients, adverse events (AEs) or intolerance to imatinib treatment required a switch to another TKI [19].

It is also noteworthy that the percentage of optimal response to imatinib may vary according to the initial clinical and hematological features of CML patients and their initial risk category, as established by the Sokal’s, Euro, and also the more recent EUTOS score [27–29]. In the IRIS study, patients with low-, intermediate-, or high-risk Sokal’s score showed significantly different response rates as 5-year CCyR (89 %, 82 %, and 69 %, respectively, P < 0.001) and progression to advanced disease (3 %, 8 %, and 17 %, respectively, P = 0.002) [17].

Based on all these considerations, several clinical trials aiming to improve first-line treatment of patients with chronic phase CML have been performed or are currently under recruiting. In particular, the first-line administration of the second-generation TKIs, or modified imatinib-based regimens, as higher dosages of imatinib from the start, or combinations of imatinib with other drugs, namely, interferon-alpha (IFN-α), have been evaluated. Currently, the only approved and registered first-line therapy regiments are the use of the second-generation TKIs nilotinib, at the dosage of 300 mg BID, and of dasatinib 100 mg OD. These two novel strategies are also included in the ELN and NCCN recommendations, whereas the other two quoted options still remain investigational [4]. Since CP-CML patients show a very long survival, very long follow-ups are mandatory to assess whether these alternative treatments promote a better OS. Meanwhile, the investigation of the efficacy of these regimens relies on important surrogate markers, as the rates of CCyR, MMR, MR4, and MR4.5; the early molecular response (EMR); and the more traditional event-free survival (EFS) and progression-free survival (PFS). However, it is important to consider that the methods to asses and to report the rate of responses can sometimes vary and that the definitions of the EFS and PFS may change substantially according to the protocol in different trials. This situation can therefore introduce bias that render difficult to compare the results [30, 31]. Considering this potential limitation, we will now review the currently available main treatment options to imatinib 400 mg OD as first-line therapy for CP-CML patients.

Following the success of imatinib, three different more potent second-generation BCR-ABL inhibitors have been developed and tested as first-line therapy to overcome the residual resistance observed in imatinib-treated patients and to further improve the outcome of CP-CML patients [32]. These TKI drugs were already approved as second-line therapy for imatinib-intolerant or imatinib-resistant patients, namely, dasatinib (Sprycel, Bristol-Myers Squibb) [33], a dual BCR-ABL and SRC inhibitor; nilotinib (Tasigna, Novartis) [34], a potent and more selective BCR-ABL inhibitor; and bosutinib (Bosulif, Pfizer) another potent dual BCR-ABL and SRC inhibitor [35].

When administrated as second-line therapy, all these drugs showed a substantially good toxicity profile and were able to induce a CCyR rate of 40–50 % in patients with primary or secondary resistance to imatinib [25, 26]. Notably, these results were achieved even in those patients expressing BCR-ABL mutations able to confer resistance to imatinib, with some notable exceptions like the T315I mutation [36].

The efficacy and the toxicity of nilotinib and dasatinib as first-line therapy were initially assessed in phase 2 studies and their rather long follow-up [37–39]. In the GIMEMA CML Working Party, obtained in 73 newly diagnosed CP-CML patients treated with nilotinib 400 mg twice a day, the CCyR achievement was observed in 78 % of patients at 3 months and in 96 % at 6 months, whereas the MMR rates observed were 52 % and 66 %, respectively, at the same time points and 85 % at 12 months [37]. Similarly, in 100 newly diagnosed CML patients treated at the MD Anderson Cancer Center with nilotinib 400 mg twice daily (BID), with a median follow-up of 29 months (range 1–73), the cumulative CCyR rate was 93 %, the MMR rate was 73 %, and the CMR rate (defined according to the previous ELN criteria as undetectable hybrid transcripts with a sensitivity of at least 10^−4/−5) was 33 % [37]. At the same institution, 86 newly diagnosed patients were also treated with dasatinib 50 mg twice daily (BID) or 100 mg QD [39]. With a median follow-up of 24 months, most patients achieved a rapid CCyR (94 % at 6 months), with a cumulative CCyR ratio of 98 %. After 12 and 18 months, MMR was achieved by 71 % and 79 % of patients [39]. The toxicity profile with dasatinib was also favorable, with a better tolerability with dasatinib QD versus BID dosing.

ENESTnd is a phase 3, randomized, open-label, multicenter study comparing the efficacy and safety of nilotinib with imatinib in patients with newly diagnosed CML that has now reached the fifth year of follow-up [40, 41]. The trial included 846 patients randomly assigned 1:1:1 to nilotinib 300 mg BID (n = 282), nilotinib 400 mg BID (n = 281), or imatinib 400 mg/day (n = 283). The primary end point was the achievement of MMR at 12 months. Patients were stratified by Sokal’s score with equal distributions of low-, intermediate-, and high-risk Sokal’s scores in each arm of the trial. Efficacy results were presented in the intent-to-treat (ITT) population. The MMR rate at 12 months was significantly higher for nilotinib 300 mg BID (44 %, P < .0001) and nilotinib 400 mg BID (43 %, P < .0001) than for imatinib (22 %). As this was the primary end point of the study, nilotinib 300 mg BID was approved by the FDA and EMA as first-line therapy. Responses were rapidly achieved with nilotinib, with 6-month MMR rates of 33 %, 30 %, and 12 % for nilotinib 300 mg BID, nilotinib 400 mg BID, and imatinib, respectively. Nilotinib treatment was also associated with less AP/BC progressions than in imatinib-treated arm [39]. After a minimum follow-up of 5 years, rates of MMR and MR4.5 continue to be significantly higher in both nilotinib arm and the imatinib arm (MMR 77 and 77.2 % vs. 60 % and MR4.5 53.5 and 52.3 % vs. 31.4 %), with more than half of the nilotinib-treated patients achieving MR4.5 by 5 years [40]. Notably, the rates of freedom from progression to AP/BC in nilotinib-treated patients remain statistically higher (96.3 % and 97.8 % for nilotinib vs. 92.1 % imatinib), when considering progression events occurring during treatment and after treatment discontinuation. However, although estimated rates of OS are higher in both nilotinib arms versus the imatinib arm (93.7 % nilotinib 300 mg BID, 96.2 % nilotinib 400 mg BID, and 91.7 % imatinib), the moment the difference is barely statistically significant only for nilotinib 400 mg BID versus imatinib. The frequency of adverse events (AEs) causing treatment discontinuation was lowest in the nilotinib 300 mg BID arm (12.2 %), followed by the imatinib arm (13.9 %) and the nilotinib 400 mg BID arm (19.9 %) [40]. However, the occurrence of cardiovascular events, which have been frequently reported in association with nilotinib therapy, has been observed in both nilotinib and imatinib arms, although these events (including all definitions of different gravity and also cerebrovascular events and peripheral arterial disease (PAD)) are definitely more frequent in the 400 mg BID arm than in the 300 mg BID arm (7.5 % in the nilotinib 300 mg BID arm, 13.4 % in the nilotinib 400 mg BID arm vs. 91.7 % in the imatinib arm) [40]. In conclusion, the 5-year follow-up data confirm the sustained efficacy of frontline nilotinib over imatinib as frontline therapy including the achievement of earlier and deeper molecular responses and increased freedom from progression to AP/BC. The relevance of these responses relies also on the option for some patients reaching the MR4.5 to discontinue the therapy without recurrence of the disease at least for a relevant period of time [19]. It is also relevant that, comparing only nilotinib 300 mg BID and imatinib 400 mg OD at 3 months, 91 % of patients in the nilotinib arm versus 67 % in the imatinib arm achieved BCR-ABL^IS transcript levels ≤10 % and 56 % and only 16 % of patients achieved already BCR-ABL^IS transcript levels ≤1 % [15]. This rapid decrease of the leukemic mass correlates with progression to AP/BC and with OS in both treatment arms. Among patients who achieved ≤10 % BCR-ABL^IS at 3 months, only three progressed on treatment whereas nine of 111 patients who achieved >10 % at 3 months progressed. These results clearly show the relevance to evaluate early molecular response at 3 months [15].

DASISION is a phase 3, randomized, open-label, multicenter study comparing the efficacy and safety of dasatinib 100 mg OD as first-line therapy with respect to that of imatinib [42]. Also this study has now achieved a minimum follow-up of 5 years [43]. All newly diagnosed CP-CML were stratified according to the Euro score and randomly assigned to dasatinib 100 mg/day or imatinib 400 mg/day. The primary end point was the CCyR by 12 months. After 1 year of treatment, dasatinib (83 %, P < .001) induces a higher response compared to imatinib (72 %), allowing this drug to be approved as first-line therapy by FDA and EMA. The best cumulative MMR rate by 12 months was also significantly higher for dasatinib (46 %, P < .0001) than for imatinib (28 %) [42]. Fewer progressions to accelerated phase or blast crisis (AP/BC) with dasatinib (1.9 %) than with imatinib (3.5 %) were already observed in the first report of these data [42]. Currently, after 5 years of follow-up, the rates of molecular response continue to be higher for dasatinib compared with imatinib (rates of MMR 76 % vs. 64 %, P = .002, and rates of MR4.5 42 % vs. 33 %, P = .025). As a consequence of these observations, the AP/BP progressions were lower with dasatinib (n = 12/259; 4.6 %) compared with imatinib (n = 19/260; 7.3 %). However 5-year PFS and OS rates were similar across treatment arms (PFS 85 % dasatinib, 86 % imatinib; OS 91 % dasatinib, 90 % imatinib) [43]. A higher proportion of patients on dasatinib achieved BCR-ABL^IS ≤10 % at 3 months (84 %) compared with those on imatinib (64 %). Patients who achieved BCR-ABL^IS ≤10 % versus >10 % at 3 months showed improved PFS, OS, and lower rates of transformation to AP/BP (PFS 89 % vs. 72 %, P = .0014; OS 94 % vs. 81 %, P = .0028; transformation n = 6/198 [3 %] vs. n = 5/37 [14 %]) and imatinib (PFS 93 % vs. 72 %, P < .0001; OS 95 % vs. 81 %, P = .0003; transformation n = 5/154 3 % vs. n = 13/85, 15 %) [16]. Regarding the total incidence of AEs, dasatinib was associated with pleural effusion in 29 %, but most cases were grade 1 or 2 (67 out of 74), and discontinuation of dasatinib due to pleural effusion occurred in only 15 patients (6 % overall and 20 % of patients who experienced a pleural effusion). Arterial ischemic events were uncommon, occurring in 12 patients (5 %) on dasatinib and 6 patients (2 %) on imatinib [43]. More recently, however, one investigator-initiated study comparing dasatinib 100 mg OD with imatinib 400 mg OD, although showing that the proportion of patients achieving CCyR was superior with dasatinib (84 % vs. 69 %) as well as the 12-month molecular responses (MMR 53 % vs. 35 %, P = 0.049; MR4 25 % vs. 10 %, P = 0.038), did not show any advantage in terms PFS as well as in terms of OS [44].

Finally, BELA is a phase 3 multicenter study comparing the efficacy and safety of bosutinib 500 mg OD with that of imatinib 400 mg OD [45]. The primary end point of this study was the CCyR by 12 months. Notably, bosutinib did not induce higher CCyR by 12 months (70 %), when compared with imatinib (68 %), and this did not allow bosutinib to be approved as first-line therapy. The high rate of bosutinib discontinuation mainly due to nonhematologic drug-related AEs may have jeopardized these results. In particular, the discontinuation rates were 19 % in the bosutinib arm with respect to 5 % in the imatinib arm and were associated with diarrhea on bosutinib. However, MMR rates by 12 months were significantly higher for bosutinib (39 % bosutinib vs. 26 % imatinib, P = 0.002), and there were numerically fewer progressions to AP/BC with bosutinib (2 %) than with imatinib (4 %) [45].

In conclusion, the first-line therapy with second-generation TKIs, due to their strongest inhibitory activity of the BCR-ABL kinase, displays superior features when compared to imatinib 400 mg OD. This is revealed by a faster time to cytogenetic and molecular responses, with more patients achieving BCR-ABL^IS ≤10 % at 3 months and by sustained higher cumulative responses, particularly by higher rates of very deep molecular responses like MR4 and MR4.5. The immediate clinical advantage of their use as frontline therapy could be represented by a lower rate of transformation, whereas on a longer run, the advantage could be represented by a faster achievement of conditions allowing to reach and maintain a TFR state. However, a relevant observation is that after a 5-year follow-up, the OS rates are not statistically different with respect to imatinib while, on the contrary, few long-term toxicity effects, like a higher rate of cardiovascular events, could raise concerns for their use, particularly in some categories of patients [46].

The reasons underlying the decision of changing TKI therapy may be different. Besides imatinib treatment failure, which is associated with a high risk of progression and death, in general, 10–12 % of patients may show adverse events (AEs) and become intolerant to treatment with a given TKI and switch treatment with another drug [10]. Notably, after 8 years of TKI treatment, only 55–60 % of patients who started with imatinib are still on treatment with this drug [21]. Interestingly, a similar discontinuation rate was observed among patients who started therapy with second-generation TKIs as first-line therapy [40, 42]. Therefore the availability of several TKIs with different characteristics and different toxicity profiles, which are registered as first-, second-, or third-line therapy, represents a great step forward in CML treatment.