Treatment initiation and first line tyrosine kinase inhibitor (TKI) selection for older patients with newly diagnosed chronic myeloid leukemia (CML) in the United States (US) before and after the introduction of generic imatinib (GE-IMA) Free

RW, JMS co-first authors; XM, NP co-senior authors

Background The first-generation TKI imatinib (IMA) is the most prescribed TKI for older patients (pts) with CML, and is preferred for pts with comorbidities such as cardiovascular disease and its risk factors. While second-generation TKIs have seen increased use over time in this population, (Shallis et al. Ther Adv in Hematol 2021), older adults remain undertreated with only 68-74% initiating TKI therapy within 3 to 6 months of diagnosis (Winn et al. JCO 2016; Shallis et al. ASH 2020). The recent introduction of GE-IMA in February 2016 has substantially lowered costs, potentially improving access for older adults with CML.

The impact of GE-IMA availability on treatment initiation and first-line (1L) TKI selection in older adults with newly diagnosed CML remains unknown.

Methods We conducted a retrospective cohort study of older pts with CML in the Surveillance, Epidemiology, and End Results Medicare-linked database. Pts were required to 1) be diagnosed from 2007-2019, 2) be ≥ 66 years (yrs), 3) have continuous Medicare Part A/B enrollment from 1 year before CML diagnosis through the end of follow-up, 4) have continuous Part D coverage from 6 months before diagnosis to end of follow-up, 5) not have received a TKI within 6 months before diagnosis, and 6) followed for ≥12 months from diagnosis. We divided the pts based on the year of CML diagnosis into two eras: before GE-IMA approval (2007-2015, pre-GE) and after (2016-2019, post-GE). We used multivariable Cox proportional hazards models to examine associations between variables and TKI treatment initiation.

Results We identified 2016 pts with CML diagnosed between 2007-2019 with a median age of diagnosis of 75 (interquartile range [IQR]: 70-81) yrs. 272 (13.5%) pts were 85 yrs or older. The majority of pts were white (81.8%), male (50.2%), married (56.5%), resided in metropolitan areas (83.2%), had ≥ 1comorbid condition (72.6%), and were not frail (66.1%). 383 (19%) pts were classified as receiving low-income subsidies.

Overall, 1438 (71.3%) pts initiated TKI therapy within 12 months of diagnosis with a median time to initiation (TTI) of 41 (IQR: 27-66) days. In the pre-GE era, 895 of 1153 (77.6%) pts initiated a TKI compared to only 543 of 863 (62.9%) pts in the post-GE era (p<0.01). The median TTI was similar in both time periods (p=0.23).

1L IMA was prescribed in 871 (60.6%) pts, with a lower proportion in the post-GE era compared to the pre-GE era (57.3% vs. 62.6%, p=0.05). The median TTI of IMA was 39 (IQR: 26-62) days, with no difference between both eras (p=0.33) The use of GE-IMA among IMA users in 1L increased over time from 65.6% (2016) to 98.8% (2019).

Among 567 (39.4%) pts who received non-IMA TKIs in 1L, the majority (98.2%) were prescribed either dasatinib (359, 63.3%) or nilotinib (198, 34.9%). The use of non-IMA TKIs in 1L increased from 37.4% in pre-GE era to 42.7% in post-GE era (p=0.05). The use of non-IMA TKIs increased from 1.6% in 2009 to 51.6% in 2015 (p for trend <0.01), but remained stable at approximately 40% per year in post-GE era (p for trend=0.72)

In multivariable analysis, post-GE era (2016-2019 vs. 2007-2015; hazard ratio [HR] 1.18; confidence interval [CI]: 1.05-1.31; p<.01) and receiving low-income subsidy (HR 1.64; CI: 1.42-1.90; p<.01) were associated with higher likelihood of TKI initiation within 12 months of diagnosis. Being older (age 85+ vs. 66-74 yrs; HR 0.79; CI: 0.69-0.3 p=0.01), unmarried (HR 0.87; CI 0.77-0.98; p=0.02) and from a Southern (vs. Northeast) region of the US (HR 0.78; CI: 0.67-0.91; p<0.01) were associated with a lower likelihood of TKI initiation, which was not affected by sex, race, frailty, comorbidity score, socioeconomic status, or living in a metropolitan area.

Conclusions Although the majority of older pts used IMA as 1L treatment in both eras, the introduction of GE-IMA did not lead to an increase of IMA utilization in the 1L setting. Moreover, treatment initiation remained suboptimal, with nearly 30% of pts not receiving any TKI within 12 months of CML diagnosis. Higher likelihood of timely TKI initiation among pts receiving low-income subsidy suggests that out-of-pocket cost remains a barrier to starting TKI treatment in the GE-IMA era.

This research was supported by the Frederick A. Deluca Foundation.