Imatinib compared with second‐generation tyrosine kinase‐inhibitors in persons with chronic myeloid leukemia presenting in accelerated phase

Nowadays, imatinib and second generation tyrosine kinase inhibitors (2G-TKIs) as initial therapy are used in newly-diagnosed chronic myeloid leukemia (CML) present in the accelerated phase. However, rare data compared responses and outcomes between them. Data from 430 consecutive subjects with CML presenting in accelerated phase from 37 medical centers across China from May 2009 to March 2022 were interrogated. The diagnosis of accelerated phase and TKI-dose adjustment were based on European LeukemiaNet recommendations.1 Last follow-up was in September 2022. Transformation was defined as blood or bone marrow blasts ≥30% during TKI therapy. Transformation-free survival (TFS) was calculated as the interval from starting TKI-therapy to transformation, death, or censored at a transplant or last follow-up.2 Survival was calculated as the interval from starting TKI-therapy to death from any cause or censored at a transplant or last follow-up. Cumulative incidences of complete cytogenetic response (CCyR), major molecular response (MMR), and molecular response 4.5 (MR4.5) were calculated using the Fine-Gray test considering competing events defined as switching to another TKI or therapy, transplant, or death. TFS and survival were calculated by the Kaplan–Meier method and compared by the log-rank test, subjects were censored at switching to another TKI or therapy, a transplant, or the last follow-up. Propensity-score matching (PSM) was done to adjust for differences in baseline covariates between the cohorts including sex, age, splenomegaly on physical exam, comorbidity(ies) (based on Charlson Comorbidity Index),3 WBC counts, hemoglobin concentration, platelet concentration, percentages of blood and/or bone marrow blasts and blood basophils, high-risk additional chromosomal abnormality(ies)1 in Ph-chromosome positive cells between subjects receiving initial imatinib or a 2G-TKI and balance evaluated using the standardized absolute mean difference where score <0.02 was considered balanced.4 A 2-sided p < .05 was considered significant. SPSS 22.0 (SPSS, Chicago, IL) and R version 4.0.2 (R Core Team, Vienna, Austria) were used for analysis and graphing. Patients' characteristics were displayed in Table S1. At the last follow-up, 166 subjects in the imatinib cohort (62%) and 121 in the 2G-TKI cohort (74%) remained on their first TKI (p = .02). Hundred subjects (38%) initially receiving imatinib switched to nilotinib (n = 43), dasatinib (n = 46), ponatinib (n = 1), olverembatinib (n = 1), flumatinib (n = 6), chemotherapy (n = 3) because of therapy-failure (n = 82), adverse events (AEs, n = 7) or subject and/or physician's preference (n = 11). Forty three subjects (26%) receiving initial 2G-TKI switched to imatinib (n = 5), nilotinib (n = 3), dasatinib (n = 25), ponatinib (n = 3), olverembatinib (n = 3), interferon (n = 1) or flumatinib (n = 3) because of therapy failure (n = 31), AEs (n = 8) or cost (n = 4). Grade 3/4 haematologic AEs occurred in 38 subjects in the imatinib cohort (14%) and 42 in the 2G-TKI cohort (26%) during the first 3 months of TKI therapy (p = .03). Fifty four subjects in the imatinib cohort (20%) and 54 in the 2G-TKI cohort (33%) had a dose reduction and/or -discontinuation (p = .03). Among them, 73 (68%) were because of grade 3/4 thrombocytopenia (n = 46), leukopenia (n = 10), or both (n = 17); 28 (26%), non-haematologic toxicity. With a median follow-up of 39 (IQR, 18–73) months. 382 (90%) subjects achieved complete hematologic response (CHR) in 3 months. 5-year cumulative incidences of CCyR, MMR, and MR4.5 were 89% (95% Confidential Interval [CI], 85, 92%), 81% (76, 86%), and 53% (47, 59%). In the cohort of subjects with excess blasts, basophils ≥20%, platelet <100 × 10E+9/L and those with ≥2 of these co-variates 5-year cumulative incidences of CCyR were 79% (67, 92%), 95% (92, 99%), 78% (67, 89%), 69% (41, 98%, p-value for trend = .01), cumulative incidences of MMR, 75% (63, 88%), 85% (79, 91%), 74% (59, 90%), 74% (49, 99%, p-value for trend = .13) and cumulative incidences of MR4.5, 56% (41, 72%), 56% (47, 64%), 47% (33, 62%), 34% (8, 61%, p-value for trend = .14). During follow-up 135 subjects (31%) failed ≥1 TKIs and 43 (10%) transformed to the blast phase. Twenty four subjects (6%) died of leukemia progression (n = 22) or other causes (n = 2). 5-year probabilities of TFS and survival were 88% (85, 91%) and 93% (90, 96%). In the cohort of subjects with excess blasts, basophils ≥20%, platelet <100 × 10E+9/L or ≥2 criteria, 5-year probabilities of TFS were 82% (72, 92%), 93% (90, 97%), 86% (78, 94%), 65% (37, 94%, p-value for trend = .01); survival, 92% (81, 100%), 95% (92, 98%), 93% (86, 99%), 74% (46, 100%, p-value for trend = .19). Results of uni-variable and multi-variable analyses were displayed in Tables S2–S7. In the total population and cohort with basophils >20% receiving 2G-TKI was associated with higher cumulative incidences of CCyR, MMR, and MR4.5, whereas in the cohort with platelet <100 × 10E+9/L receiving 2G-TKI was associated with worse TFS. We did not analyze subjects with ≥2 criteria because of too few subjects. We used 1:2 PSM to adjust for differences in baseline co-variates in the two cohorts. Three hundred and thirty one matches were identified in the imatinib (n = 198; 60%) and 2G-TKI (n = 133; 40%) cohorts (Table S8). There was no significant difference in CHR (both 92% vs. 88%, p = .23). Median follow-up of subjects receiving imatinib and 2G-TKI were 53 (IQR, 26–94) and 27 (IQR, 13–54) months (p < .001). 29 (15%) and 33 (25%) subjects had a grade 3/4 haematologic AEs (p = .02), 38 (19%) and 42 (32%) underwent dose reduction or interruption (p = .01). One hundred and eighteen subjects in the imatinib cohort (60%) and 99 in the 2G-TKI cohort (74%) remained on their first TKI at the last follow-up (p = .01). Subjects initially receiving a 2G-TKI had higher cumulative incidences of CCyR (p = .01), MMR (p = .001), and MR4.5 (p = .02). There was no difference in TFS (p = .14) or survival (p = .94) between the two cohorts (Figure 1A). In 331 matched subjects, 318 subjects (96%) met 1 criterion for accelerated phase including 63 (19%) with excess blasts, 209 (63%) with basophils >20%, and 46 (14%) with platelet <100 × 10E+9/L. Thirteen subjects had ≥2 criteria, too few to analyze (Table S8). In the cohort of subjects with excess blasts, 37 subjects received imatinib and 26 received a 2G-TKI. 6 (16%) and 10 (39%) had a grade of 3/4 AEs (p = .05), 5 (14%), and 10 (39%) a dose-reduction or -interruption (p = .02), 21 (57%) and 14 (54%) remained on their initial TKI (p = .82). There were no differences in cumulative incidences of CCyR, MMR, MR4.5, TFS, and survival between the two cohorts (Figure 1B). In the cohort of subjects with basophils ≥20%, 127 subjects received imatinib and 82 received a 2G-TKI. 21 (17%) and 15 (18%) had a grade 3/4 haematologic AEs (p = .74), 29 (23%) and 21 (26%) had a dose-reduction or -interruption (p = .65), and 77 (61%) and 67 (82%) remained on their initial TKI (p = .001). Subjects receiving initial 2G-TKI therapy had higher cumulative incidences of CCyR (p = .01), MMR (p < .001), and MR4.5 (p = .01). Subjects receiving 2G-TKI had similar TFS and survival compared to those receiving imatinib (Figure 1C). In the cohort of subjects with platelet concentration <100 × 10E+9/L 27 received imatinib and 19 received a 2G-TKI. 1 (4%) and 7 (37%) had a grade 3/4 haematologic AEs (p = .01), 1 (4%) and 8 (42%) a dose-reduction or -interruption (p = .002) and 16 (59%) and 14 (74%) remained on their initial TKI (p = 0.31). There was no significant difference in cumulative incidences of CCyR, MMR and MR4.5, TFS, and survival between the two cohorts (Figure 1D). Similar results were observed in sub-cohorts with different features of accelerated phase when switching to second-line therapy were considered as competing or censoring events (Figures S1–S4). Both the 2G-TKI and imatinib cohorts included subjects receiving branded and generic drug versions. In sensitivity analyses our conclusions were unchanged (Figures S5 and S6). Our study has several limitations. First, it was retrospective. Second, data were provided by physicians instead of a study-designed instrument. Third, we did not audit centers to guarantee reporting of consecutive subjects for data accuracy. Lastly, we did not monitor TKI compliance. The recent World Health Organization (WHO) CML classification eliminates the accelerated phase as a disease phase category. We agree with this change. However, considerable data indicate it takes several years for physicians to adopt consensus recommendations (reviewed in Clinical Practice Guidelines We Can Trust5). In the interim physicians are likely to continue using ELN accelerated phase criteria in their TKI decision making. In summary, we found no advantage in outcomes using a 2G-TKI in persons with CML initially presenting in an accelerated phase compared with imatinib except for better molecular response in subjects with basophils >20%. In some instances, there was a disadvantage seemingly reflecting 2G-TKI-associated dose interruptions and/or discontinuations because of hematological AEs. Our conclusions need validation in a randomized controlled trial but may help physicians choose the best initial TKI therapy for subjects with CML initially presenting in the accelerated phase. Whether a third generation TKI might be advantageous in this setting is unknown. Qian Jiang and Xiao-jun Huang designed the study. Qian Jiang, Sen Yang, and Xiaoshuai Zhang analyzed the data. Sen Yang, Xiaoshuai Zhang, Xin Du, Chun-yan Chen, Jian-yu Weng, Jian Huang, Fei Li, Yun Zeng, Zhen Xiao, Jian-da Hu, Li-jie Yang, Zhuo-gang Liu, Guo-hui Li, Xiu-li Sun, Wei Yang, Ru Feng, Yan-qiu Han, Yu Jing, Na Xu, Xiao-li Liu, Zhen-fang Liu, Xiao-dong Wang, Shi-xin Wu, Rong Liang, Yan-li Zhang, Yun-fan Yang, Huan-ling Zhu, Ling Pan, Li Meng, Yan-hong Zhao, Hai Yi, Yi-lan Liu, Wei-hua Zhang, Yuan-jun Zheng, Ze-ping Zhou, Su-ning Chen, Hui-ying Qiu, Wei-ming Li, Zhi-lin Jia, Yan-liang Bai, Li-e Lin, Bing-cheng Liu, Chun-shui Liu, Jian-min Luo, Jun-xia Meng, Zhi-qiang Sun, Yan-qing Zhang collected the data. Qian Jiang, Sen Yang, Xiaoshuai Zhang, Robert Peter Gale, and Xiao-jun Huang prepared the typescript. All authors approved the final typescript, take responsibility for the content, and agreed to submit it for publication. The authors thank medical staff and patients. RPG acknowledges support from the National Institute of Health Research (NIHR) Biomedical Research Centre and the Ministry of Science and Technology of China (84000-51200002). Funded, in part, by the National Nature Science Foundation of China (No. 81970140). Robert Peter Gale is a consultant to BeiGene Ltd., Fusion Pharma LLC, LaJolla NanoMedical Inc., Mingsight Parmaceuticals Inc., and CStone Pharmaceuticals; advisor to Antegene Biotech LLC, Medical Director, FFF Enterprises Inc.; partner, AZAC Inc.; Board of Directors, Russian Foundation for Cancer Research Support; and Scientific Advisory Board: StemRad Ltd. All data generated during this study are included in this published article and its supplementary information files. The raw data analyzed during the current study available from the corresponding author on reasonable request. Figure S1. Comparison of responses and outcomes using PSM analyses. Switching to another TKI, transplant, or death were considered competing events for the response. Outcomes were censored at switching to another TKI, a transplant, or the last follow-up. Figure S2. Comparison of responses and outcomes in subjects with excess blasts. Switching to another TKI, transplant, or death were considered competing events for the response. Outcomes were censored at switching to another TKI, transplant, or last follow-up. Figure S3. Comparison of responses and outcomes in subjects with basophils ≥20 percent. Switching to another TKI, transplant, and death were considered competing events for the response. Outcomes were censored at switching to another TKI, a transplant, or the last follow-up. Figure S4. Comparison of responses and outcomes in the cohort with a platelet concentration <100 × 10E+9/L. Switching to another TKI, transplant, or death were considered competing events of responses; outcomes were censored at switching to another TKI, transplantation, or at last follow-up. Figure S5. Comparison of responses and outcomes in subjects receiving branded TKIs using PSM analyses. Transplant and death were considered competing events for the response. Outcomes were censored at the transplant or the last follow-up. Figure S6. Comparison of responses and outcomes in subjects receiving generic TKIs using PSM analyses. Transplant and death were considered competing events for the response. Outcomes were censored at the transplant or the last follow-up. Table S1. Clinical and lab characteristics of patients with CML in accelerated phase. Table S2. Univariate analysis of co-variates associated with responses and outcomes in the total population. Table S3. Univariable analyses of co-variates associated with responses and outcomes in subjects with excess blasts. Table S4. Univariate analyses of co-variates associated with responses and outcomes in subjects with basophils ≥20%. Table S5. Univariable analyses of co-variates associated with responses and outcomes in subjects with a platelet concentration <100 × 10E+9/L. Table S6. Co-variates associated with responses and outcomes in multi-variable analyses in total population. Table S7. Co-variates associated with responses and outcomes in subgroups by accelerated phase criterion. Table S8. Subject co-variates after propensity score matching. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.