Comparison of the Outcome between Newly Diagnosed Patients in the Accelerated Phase and Chronic Phase with Chronic Myeloid Leukemia Treated with Frontline Tyrosine Kinase Inhibitors

Background: Accelerated phase (AP) chronic myeloid leukemia (CML) has an aggressive clinical course. Some patients with CML present as AP at the time of diagnosis. Although many studies have confirmed that tyrosine kinase inhibitors (TKIs) significantly improved survival in patients with CML in the chronic phase (CP), there is limited data on the outcome of newly diagnosed CML-AP patients in the era of TKIs.

Aims: To compare progression-free survival and overall survival in CML-AP patients with those in CML-CP patients treated with frontline TKIs and identify factors associated with the outcome.

Methods: According to the definition of European LeukemiaNet 2013, data of newly diagnosed CML-AP and CML-CP patients receiving TKIs at least 3 months from 2002 to 2016 were retrospectively analyzed. Kaplan-Meier curves were used to compare the probabilities of PFS and OS. Uni- and multivariate Cox regression models were built to identify independent variables (demographics; co-morbidity; disease phase and risk, CBC, and additional cytogenetic aberrations in Ph-positive cells at diagnosis; interval from diagnosis to start TKI-therapy; initial TKI used and early treatment-response) associated with the outcome.

Results: In total, 101 CML-AP patients [with excess blasts (blasts ≥15%), n=13; with excess basophils (basophils ≥20%), n=88] and 656 CML-CP patients (Sokal low-risk, n=291; intermediate-risk, n=194; high-risk, n=111; unknown, n=60) were included. 463 (61%) were male. Median age was 39 (14-87) years. Patients received initial therapy with imatinib (n=660), nilotinib (n=85), or dasatinib (n=11) within 6 months from diagnosis. At 3 months, CML-AP patients with excess blasts (62%, 15% and 23%) and excess basophils (65%, 27% and 8%) achieved similar response (optimal, warning and failure) rates to CML-CP patients in Sokal intermediate-risk (66%, 20% and 14%) and high-risk (58%, 26% and 16%) (p=0.338), but significantly lower than those in Sokal low-risk (83%, 13% and 4%) and unknown (77%, 17% and 7%) (p<0.001). With a median follow-up of 39 months (range, 3-184 months), probabilities of PFS (p=0.014) and OS (p=0.152) at 5 years were the lowest in the CML-AP patients with excess blasts (78% and 92%), followed by CML-CP patients in Sokal high-risk (87% and 91%), intermediate-risk (92% and 95%), low-risk (95% and 97%), CML-AP patients with excess basophils (95% and 97%), and those with unknown-risk (96% and 98%). In multivariate analyses, while analyzing baseline and initial therapy variables, interval from diagnosis to starting TKI-therapy >2 months [HR 2.4 (1.1-4.9), p=0.020 and HR 2.6 (1.1-6.5), p=0.037) was associated with inferior PFS and OS; in addition, CML-CP patients in Sokal high-risk [HR 3.6 (1.1-6.5), p=0.038] was associated with inferior PFS, >60 years [HR 3.9 (1.2-13.3), p=0.027], inferior OS. While involving early therapy response, interval from diagnosis to starting TKI-therapy >2 months [HR 3.1 (1.4-6.6), p=0.009 and HR 3.2 (1.2-8.4), p=0.016] and achieving a response of "failure" at 3 months [HR 7.6 (3.7-15.6), p<0.001 and HR 8.1 (3.1-21.3), p<0.001] were variables associated with inferior PFS and OS; in addition, >60 years [HR 4.8 (1.4-16.1), p=0.010], inferior OS. There was no impact of initial TKI used (imatinib or 2^nd generation TKI) on the outcome.

Conclusions: Newly diagnosed CML-AP patients have comparable outcome to CML-CP patients treated with frontline TKIs, especially CML-AP patients with excess basophils. However, delayed therapy and poor early response to TKIs are significantly associated with unfavorable prognosis in either AP or CP patients with CML. Our data indicates timely and proper TKI-therapy and close monitoring will be a reasonable option for the treatment strategy in CML-AP patients in the era of TKIs.