The basis of the NCCN recommendation is that 2G-TKIs lead to improved molecular and cytogenetic responses in CML patients. A meta-analysis of randomized controlled trials comparing second and third-generation TKIs to imatinib showed risk ratios of CCyR (defined as the absence of Ph+ metaphases) and MMR (defined as 3-log reduction in BCR-ABL1 transcripts) at 12 months to be 1.13 and 1.50, respectively [3]. However, it is unclear whether deeper molecular and cytogenic responses translate to improved patient-centered outcomes, and recent evidence suggests the contrary. Bidikian et al. reported long-term outcomes of 131 patients who did not achieve MMR after 2 years of treatment with TKIs, finding that 10-year CML-related OS was 95% if MCyR was achieved and 80% if MCyR was not achieved [4]. MMR is a poor measure of treatment failure, as patients who fail to achieve MMR can still achieve good outcomes. There is very limited data on correlations between CCyR and MMR with OS across multiple randomized controlled trials (i.e. level-1 evidence).

Importantly, 2G-TKIs have failed to demonstrate any improvement in OS or health-related quality of life over imatinib in randomized controlled trials [3]. The ENESTnd study comparing imatinib and nilotinib reported a 10-year OS of 88.3% in the imatinib arm vs 90.3% in the nilotinib (400 mg) arm [5]. No significant difference in OS was found, even though the incidence of progression to accelerated phase/blast phase was suppressed in the nilotinib group. In the DASISION study on imatinib vs dasatinib, 5-year OS was 90.0% and 91.0% in the imatinib and dasatinib arms, respectively [3]. The BFORE trial on imatinib vs bosutinib found similar 12-month OS between treatment groups. While rates of treatment-free remission (TFR) eligibility as defined by molecular measurements by RT-PCR were higher with nilotinib in the ENESTnd study, data on overall TFR success rates was not provided. Actual rates of TFR with imatinib, nilotinib, and dasatinib have been found to be similar (~50%) in discontinuation trials [6]. Based on current data, first-line use of 2G-TKIs provides no real clinical benefit to the patient but adds significant toxicity and cost.

2G-TKIs are arguably more toxic. Specifically, they are associated with cardiovascular, pulmonary, pancreatic, and hepatic toxicities [3]. In the ENESETnd study, 10-year cumulative incidence of cardiovascular events was 24.8% in the nilotinib (300 mg bid) arm as opposed to 6.3% in the imatinib arm [7]. Nilotinib is also associated with glucose tolerance and dyslipidemia, and its use in patients with cardiovascular risk factors or diabetes requires careful consideration. In the DASISION study, dasatinib was found to be a risk factor for pleural effusion and pulmonary hypertension, and patients should be evaluated for pulmonary disease before treatment. The exclusion criteria for trials assessing the efficacy of 2G-TKIs were broader than those used for imatinib alone given their toxicity profiles [8]. Toxicity rates may thus be higher than what trials of 2G-TKIs report. Given that the median age of diagnosis of CML is greater than 60 years, many patients have comorbidities resulting in high risk of treatment-related adverse effects. Patients with comorbidities who are treated with 2G-TKIs require monitoring, resulting in additional medical expenses and time.

2G-TKIs may also be associated with higher likelihood of treatment interruptions. One study using real world data from a claims database found that 59% of patients who received 2G-TKIs had treatment interruptions compared to 45% for imatinib [9]. Similar treatment interruption rates were reported in the ENESTnd and DASISION trials. These studies could not evaluate the reasons for treatment interruption, but toxicity is likely a key factor. Minimizing 2G-TKI treatment interruptions has been shown to lead to better outcomes including greater failure-free survival.