Burden of tyrosine kinase inhibitor failure in Chinese chronic myeloid leukemia patients: a systematic literature review

We consulted the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Statement (http://www.prisma-statement.org/PRISMAStatement/) to design this SLR. In brief, we conducted a literature search of major English- and Chinese-language bibliographic databases of RWE studies from published, peer-reviewed articles concerning the disease burden and unmet medical needs of Chinese patients with CML and TKI treatment failure. Treatment failure was in turn defined as primary and secondary drug resistance or treatment intolerance associated with TKIs for CML since the approval of imatinib for treating CML in mainland China in 2002.

This SLR included real-world studies reporting: the epidemiology of CML; treatment failure and BCR-ABL1 mutations associated with the three approved TKIs for CML in China (imatinib, nilotinib and dasatinib); patient characteristics; disease progression; treatment information; treatment outcomes; HRQoL; HRU; and direct medical costs in Chinese patients with CML and TKI treatment failure.

To our knowledge, the present SLR is the first comprehensive evaluation of RWE related to disease burden and unmet medical needs in Chinese patients with CML and TKI treatment failure. Encompassing 20 years of RWE data, our literature search ensured inclusion of relevant evidence released after approval of the first TKI for CML in China (imatinib). Although the findings of stable age-standardized incidence and sharply decreased age-standardized mortality rates are encouraging, our SLR also adds substantial dimensions to the known burdens of TKI failure in CML. These include: an elevated risk of developing treatment failure associated with 1L and subsequent-line treatments; a considerable prevalence of BCR-ABL1 kinase domain mutations (especially the T315I gatekeeper); and rapid disease progression. Patients with more advanced CML and/or those with greater numbers of TKI treatment lines (secondary to treatment resistance and failure) experienced: compromised treatment outcomes, including lower molecular and cytogenetic responses and survival; marked and clinically meaningful decrements in HRQoL; and sharply increased resource utilization and costs of care.

Like epidemiological patterns in other countries, China has had a stable age-standardized incidence and decreasing age-standardized mortality of CML, largely reflecting the advent and era of TKI therapy. Although imatinib was approved for the treatment of CML in 2002, Chinese patients with CML had limited access to TKIs until the recent approval of a reimbursement schedule and substantial price decreases for imatinib via a volume-based procurement policy [112]. Therefore a great expansion of patient access to imatinib is expected, which could further decrease CML-related mortality in China. One study forecast that the US prevalence of CML will increase from the year 2000 to a plateau by 2050 because of TKI treatments [113]. Because China has a similar crude incidence rate of CML to the USA (1–1.3 per 100,000) [114], the prevalence of CML in China could follow a similar trend. Future epidemiological studies are needed to confirm the trend of CML prevalence and provide robust evidence predicting the CML population size in China.

CML is difficult to cure, and the present review shows that imatinib was associated with the greatest risk of treatment failure (~20% per year) among the three approved TKIs in China. Although patients with failed imatinib treatment could be switched to nilotinib or dasatinib for continued treatment, the present review shows that treatment with 2G TKIs after imatinib treatment failure was associated with even greater risks of drug resistance and treatment intolerance. Consequently, such patients might face a greater risk of treatment failure. Absent an effective solution for TKI treatment failure, patients with such unmet needs could outnumber patients with controlled disease and become the main driver of CML-related burden in China.

TKI treatment failure in CML constitutes a formidable public health complication that can amplify the disease burden in many ways. For example, the pooled median age of Chinese patients with failed TKI treatment in our included studies (38.6 years) was approximately 10 years younger than that of patients with CML receiving imatinib as front-line treatment in China (49 years) [115]. This finding corroborated the relationship between age and risk of disease relapse in the ISAV study [116], which reported that patients younger than 45 years had a significantly greater rate of CML relapse (95%) compared with those aged between 45 and 65 (42%) or older than 65 (33%). There is no conclusive explanation for the relationship between age and disease relapse associated with imatinib; nevertheless, findings from the ISAV study suggested that younger age in patients with TKI treatment failure could substantially worsen the disease burden by extending disease duration. The proportion of high prognostic risk in patients with TKI treatment failure in our review (42.5%) was also approximately three-times that in previously untreated patients (13%) [117]. This relationship indicates that patients with failed TKI treatment had more rapid progression to advanced disease than those with previously untreated CML, which was associated with poor survival prognosis, decreased HRQoL and increased HRU and medical costs, based on evidence reported in the present SLR.

According to our present findings, the real-world risk of TKI treatment failure in Chinese patients with CML was much greater than was reported in the randomized trial setting. Our SLR showed that the pooled risk of treatment failure associated with 1L imatinib was approximately 20% per year, which is three- to four-times greater than those reported in the DASISION trial (4.8% per year) [118] and the IRIS trial (7.0% per year) [119]. Our review also found that the risk of treatment failure associated with 1L nilotinib was approximately 25% of that for imatinib but approached 30% for 2L nilotinib after failing imatinib. Given that more than 50% of patients with failed imatinib treatment subsequently received nilotinib, Chinese populations with CML and TKI treatment failure could be much larger than expected. Pooled evidence from the studies included in the present SLR confirmed that drug resistance was the predominant reason for treatment failure associated with the three approved TKIs in Chinese patients. Correlations between TKI resistance and BCR-ABL1 kinase domain gene mutations have been well established [120]. Consequently, it was not surprising to find an elevated risk of drug resistance in our studies, in which 51.6% of Chinese patients with failed TKI treatment had at least one BCR-ABL1 kinase domain mutation. Of these, T315I was the most prevalent, occurring in 10.6% of patients with TKI treatment failure. In addition, the risk of developing T315I mutations in the 2L treatment setting was nearly three-times greater than that associated with 1L imatinib. Because the T315I mutation confers resistance against both nilotinib and dasatinib [8], these genetic variants, as well as compound mutations (some including T315I), might contribute to a greater burden of TKI treatment failure in the future.

Treatment outcomes, including cytogenetic and molecular responses (and survival), were worse in patients with CML and failed TKI treatment. Nearly 80% of Chinese patients with CML were switched to a 2G TKI (mainly nilotinib) after imatinib treatment failure. Compared with the 1-year CCyR (80.1%) and MMR (44.0%) rates reported for 1L nilotinib in patients with CP-CML in the ENESTnd study [121], real-world treatment responses with 12-month nilotinib in Chinese patients with failed imatinib treatment (CCyR: 51.5%; MMR: 39.4%) were approximately 10–30% lower. Similarly, the 1-year CCyR rate with dasatinib in Chinese patients with failed imatinib treatment (47.7%) was nearly 60% of that reported for 1L dasatinib (83.4% in the DASISION study [118]). Treatment responses of the two TKIs were also decreased in the 3L setting and in patients with advanced-phase CML. A similar trend was reported for poorer survival outcomes associated with nilotinib in patients with advanced-phase CML that failed imatinib treatment.

Safety is a major potential concern when administering two 2G TKIs in patients with a history of treatment failure. According to pooled characteristics in the present SLR, more than 30% of patients with failed TKI treatment had at least one comorbidity, including approximately 10% with cardiovascular disease, which increases the relative risk of recurrent (vs initial) events with nilotinib or dasatinib [122]. Given the relatively young age of the patient population in the present SLR, safety issues might be more meaningful in more heterogeneous populations including older patients, who are more likely to have treatment-limiting chronic comorbidities. In addition, there is a high incidence of serious hematological events associated with nilotinib which can be life threatening and require hospital care [123]. Thus the two approved 2G TKIs might not completely meet the medical needs of patients with TKI treatment failure. In the event of disease relapse after treatment with nilotinib or dasatinib, HSCT is likely the only remaining treatment option. However, HSCT is associated with clinical issues and administrative hurdles such as strict eligibility requirements, prohibitively high costs, considerable risk of serious complications, and low success rate in patients with AP- or BP-CML [124].

Emerging challenges of TKI treatment failure have opened avenues for pharmacological innovation. For example, in 2015 the US FDA approved the third-generation TKI ponatinib for patients with CML and failed TKI treatment because of its marked antileukemic efficacy in patients with heavily pretreated CML, including those with T315I mutations. According to the PACE trial [125], the 1-year CCyR reported for ponatinib in patients with CP-CML that failed 2G TKIs was 46%, which was approximately 50% less in patients with AP- or BP-CML (vs CP-CML). Compared with allogeneic HSCT, ponatinib treatment was associated with survival benefits in patients with T315I-mutated CP-CML (4-year OS rate: 72.7 vs 55.8%) but not T315I-mutated AP-CML or BP-CML [126]. The promising effects of ponatinib have been tempered somewhat by reports of cardiotoxicity, arterial occlusive events, heart failure, venous thromboembolism and hepatotoxicity (along with other issues such as pancreatitis), which have resulted in black box warnings in the USA and other package labeling [127]. Ponatinib is the most cardiotoxic of 23 TKIs, with a potential mechanism of stimulating cardiomyocyte apoptosis [128–130]. In the PACE trial [125], 37% of patients receiving ponatinib had thrombocytopenia, 19% serious arterial and 5% venous occlusive adverse events. Ponatinib has not been approved in mainland China. With better understanding of the impact of patient characteristics (age, prognostic risk, comorbidity), disease molecular profiles and TKI toxicity profiles on the risk of TKI treatment failure, tailoring TKI treatments according to these factors could be another potential solution to control the emerging disease burden associated with TKI treatment failure in Chinese CML patients [131].

The present SLR identified important knowledge gaps that need to be bridged to better understand the disease burden and unmet medical needs of patients with CML failing TKI treatments in China. First, there is an absence of definitive quantitative evidence to support the premise that the prevalence of CML and TKI treatment failure will increase in China. It is therefore difficult to estimate the size of the future population of patients with failed TKI treatment, which subsequently complicates the decision-making process for reimbursement. Second, much of the evidence in the present SLR was based on 1L imatinib treatment failure; therefore, disease progression, risk of gene mutation, treatment outcomes and treatment intolerance in patients with CML that failed nilotinib and dasatinib treatments remain unclear. On the basis of limited evidence, TKI treatment failure in the 2L and 3L settings could inflict a heavier disease burden than 1L failure. Third, evidence reported in the present SLR should be considered in the context of certain limitations. Some findings, such as risks of BCR-ABL1 gene mutations with the three TKIs and treatment outcomes of the two 2G TKIs in patients with failed imatinib treatment were based on a single study with a limited sample size. Potential selection bias and uncertainty could decrease the generalizability of this evidence. Resistance to TKIs, as defined in this SLR, included both primary and secondary (acquired) resistance because most studies did not classify the outcome in this manner. Consequently, the pooled annual risk of TKI resistance could be overestimated. The included studies did not follow up with patients for sufficient durations to record survival outcomes. The present SLR used reported short-term survival data to extrapolate median outcomes using the exponential survival function assumption; thus these pooled survival outcomes need to be validated in future research. In addition, future studies are needed to fill evidence gaps and improve the robustness of the evidence for HRQoL and costs. Because Chinese publications were the main evidence source for the present SLR, the reported evidence related to treatment outcomes could be heavily biased by study quality. Finally, the common limitations associated with the nature of real-world studies could introduce bias in the reported evidence. These limitations and evidence gaps reflect a need for a representative CML patient registry to collect further, longitudinal real-world data on patients and their disease management, hence improving the validity and reliability of RWE for CML in China. Under whatever limitations, this review has comprehensively characterized the emerging disease burden and unmet medical needs in Chinese patients with CML and TKI treatment failure. Further, the synthesized evidence in this review can support future health economic evaluations for the coming third-generation TKIs which can effectively treat CML patients who failed with 1G and 2G TKIs.

This SLR of RWE shows that the disease burden of TKI treatment failure in Chinese patients with CML is substantial and poised to burgeon because of: a potentially longer disease course associated with relatively young patient age at disease onset; a high prevalence of BCR-ABL1 kinase domain mutations, which can confer TKI resistance and erode treatment outcomes; more rapid disease progression; limited treatment effectiveness associated with 2G TKIs in patients with TKI treatment failure; and the impaired HRQoL and increased HRU and medical costs associated with disease resistance and progression. New and novel, effective, well-tolerated, quality of life-enhancing and affordable treatments targeting TKI-resistant gene mutations, such as T315I and compound mutations, are urgently needed to address the unmet medical needs of Chinese patients with CML and TKI treatment failure.

YZ and W Chen were responsible for study conception, planning and design. Y Zhang was responsible for the planning and design of the systematic literature review. W Chen was responsible for data acquisition. All authors participated in data interpretation, drafting the manuscript (and/or making substantive intellectual contributions), reviewing the final version and agreeing to be accountable for the study and its findings.

The authors would like to thank H Huang and L Tan for their work on the literature search, study eligibility assessment, study quality assessment and data extraction in this study. S Gutkin and N Ogba, with Ascentage, provided substantive input in manuscript research and preparation.

This study was sponsored by Ascentage Pharma Group Corp., Ltd. (Hong Kong), which contributed to the study design, interpretation of data and the decision to submit the article for publication. W Chen is an employee of Normin Health Consulting, Ltd., which received consulting fees from Ascentage Pharma Group, Inc. A Pathak, Y Yang, D Yang and Y Zhang are employees of Ascentage Pharma Group, Inc. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No funded writing assistance was used in the creation of this manuscript.

Extracted data from the included studies can be accessed by contacting the corresponding author.