Abstract
Objective: To explore the impacts of socio-demographic and clinical co-variates on treatment responses and outcomes in patients with chronic myeloid leukemia in the chronic phase (CML-CP) receiving tyrosine kinase inhibitor (TKI) and identified the predictive models for them. Methods: Data of newly diagnosed adult patients with CML-CP receiving first-line TKI and having complete socio-demographic data and clinical information were reviewed. Cox model was used to identify the independent variables associated with complete cytogenetic response (CCyR) , major molecular response (MMR) , molecular response 4 (MR(4)) and molecular response 4.5 (MR(4.5)) , as well as failure-free survival (FFS) , progression-free survival (PFS) , overall survival (OS) and CML-related OS. Results: A total of 1414 CML-CP patients treated with first-line imatinib (n=1176) , nilotinib (n=170) or dasatinib (n=68) were reviewed. Median age was 40 (18-83) years and 873 patients (61.7% ) were males. Result of the multivariate analysis showed that lower educational level (P<0.001-0.070) and EUTOS long-term survival intermediate or high-risk (P<0.001-0.009) were significantly associated with lower cumulative incidences of CCyR, MMR, MR(4) and MR(4.5), as well as the inferior FFS, PFS, OS and CML-related OS. In addition, those who were males, from rural households, had white blood cells (WBC) ≥120×10(9)/L, hemoglobin (HGB) <115 g/L and treated with first-line imatinib had significantly lower cumulative incidences of cytogenetic and/or molecular responses. Being single, divorced or widowed, having, rural household registration, WBC≥120×10(9)/L, HGB<15 g/L, and comorbidity (ies) was significantly associated with inferior FFS, PFS, OS, and/or CML-related OS. Thereafter, the patients were classified into several subgroups using the socio-demographic characteristics and clinical variables by cytogenetic and molecular responses, treatment failure and disease progression, as well as overall survival and CML-related OS, respectively. There were significant differences in treatment responses and outcomes among the subgroups (P<0.001) . Conclusion: Except for clinical co-variates, socio-demographic co-variates significantly correlated with TKI treatment responses and outcomes in CML-CP patients. Models established by the combination of independent socio-demographic and clinical co-variates could effectively predict the responses and outcome.